1
00:00:04,160 --> 00:00:07,160
Speaker 1: Get in touch with technology with tech Stuff from how

2
00:00:07,240 --> 00:00:14,160
Speaker 1: stuff Works dot com. Hey there, and welcome to tech Stuff.

3
00:00:14,200 --> 00:00:17,560
Speaker 1: I'm your host, Jonathan Strickland. I'm an executive producer with

4
00:00:17,600 --> 00:00:20,160
Speaker 1: How Stuff Works and the Love all Things Tech Today.

5
00:00:20,200 --> 00:00:22,880
Speaker 1: Probably sounds a little different from my normal episodes because

6
00:00:22,880 --> 00:00:25,639
Speaker 1: I'm actually recording this while I'm in a hotel room

7
00:00:25,680 --> 00:00:28,639
Speaker 1: in Las Vegas. I'm attending yet another tech conference. But

8
00:00:28,760 --> 00:00:33,000
Speaker 1: this episode is more in line with our normally scheduled episodes.

9
00:00:33,320 --> 00:00:37,440
Speaker 1: This one is from a listener request, and the listener

10
00:00:37,479 --> 00:00:41,279
Speaker 1: in this case wished to remain anonymous, and so I

11
00:00:41,360 --> 00:00:46,640
Speaker 1: am going to to completely comply with his or her wishes,

12
00:00:47,200 --> 00:00:51,400
Speaker 1: but wanted to know more about end to end encryption,

13
00:00:51,680 --> 00:00:53,960
Speaker 1: which has been the news a lot, particularly in the UK.

14
00:00:54,960 --> 00:00:59,360
Speaker 1: End to end encryption is a means of sending information securely,

15
00:00:59,440 --> 00:01:03,880
Speaker 1: and it has positive aspects and depending upon who you are,

16
00:01:03,960 --> 00:01:06,440
Speaker 1: negative aspects. We're going to cover all of that, explain

17
00:01:06,520 --> 00:01:09,280
Speaker 1: what end and encryption is and how it works. It's

18
00:01:09,280 --> 00:01:11,440
Speaker 1: been the news for a couple of years for multiple reasons.

19
00:01:11,520 --> 00:01:14,000
Speaker 1: One is that some folks and government agencies have made

20
00:01:14,000 --> 00:01:16,399
Speaker 1: a stink about the technology because they say it gives

21
00:01:16,480 --> 00:01:20,520
Speaker 1: bad actors like criminals and terrorists, the ability to communicate

22
00:01:20,560 --> 00:01:24,200
Speaker 1: secretly with one another. That, in a sense is true,

23
00:01:24,400 --> 00:01:28,120
Speaker 1: but I think it oversimplifies the issue. The tech actually

24
00:01:28,160 --> 00:01:32,800
Speaker 1: allows any two people to communicate secretly with one another. So,

25
00:01:32,840 --> 00:01:34,440
Speaker 1: in other words, it's a tool that can be used

26
00:01:34,440 --> 00:01:38,360
Speaker 1: by people who are bad, but in itself is not

27
00:01:38,440 --> 00:01:41,080
Speaker 1: a bad tool. It's a tool that has tons of

28
00:01:41,200 --> 00:01:45,679
Speaker 1: legitimate uses that have nothing to do with terrorism or crime.

29
00:01:45,840 --> 00:01:48,560
Speaker 1: So as an example, let's say you are running a

30
00:01:48,640 --> 00:01:52,000
Speaker 1: record label and it's your job to sign new acts

31
00:01:52,040 --> 00:01:54,520
Speaker 1: to this record label. In the old days, you would

32
00:01:54,520 --> 00:01:57,200
Speaker 1: do all this in person, but today you can do

33
00:01:57,440 --> 00:01:59,640
Speaker 1: a lot of your business, practically all of your business

34
00:01:59,640 --> 00:02:02,240
Speaker 1: over the internet. So you can negotiate with the act.

35
00:02:02,320 --> 00:02:04,240
Speaker 1: You can come to an agreement on terms. You can

36
00:02:04,320 --> 00:02:06,480
Speaker 1: draw up a contract, and you can send it over

37
00:02:06,520 --> 00:02:08,680
Speaker 1: for them to sign, and then send it back for

38
00:02:08,720 --> 00:02:11,560
Speaker 1: you to countersign. But you want all this information to

39
00:02:11,600 --> 00:02:15,960
Speaker 1: remain confidential. The terms of one agreement with one act

40
00:02:16,120 --> 00:02:18,800
Speaker 1: could be very different from the terms you set with

41
00:02:18,840 --> 00:02:21,959
Speaker 1: another act. To lure a really big act to your label,

42
00:02:22,440 --> 00:02:26,240
Speaker 1: you might have to include more attractive terms for that act,

43
00:02:26,320 --> 00:02:30,280
Speaker 1: like better percentages, But you don't necessarily want to do

44
00:02:30,360 --> 00:02:33,720
Speaker 1: that with every act you're signing. It might not merit it.

45
00:02:33,919 --> 00:02:37,679
Speaker 1: In some cases, the amount of investment a label might

46
00:02:37,720 --> 00:02:42,000
Speaker 1: make in an unproven act could end up being unrewarded downeline.

47
00:02:42,080 --> 00:02:46,560
Speaker 1: So you would rather have the details of contracts remain

48
00:02:46,919 --> 00:02:50,040
Speaker 1: secure and not get leaked to the general public, because

49
00:02:50,080 --> 00:02:53,440
Speaker 1: then your negotiation tools are made plain for everyone to see.

50
00:02:53,760 --> 00:02:56,639
Speaker 1: It opens up the opportunity for various acts to say, hey,

51
00:02:56,639 --> 00:02:59,800
Speaker 1: how come they got such a sweet deal and we didn't. Uh,

52
00:03:00,040 --> 00:03:01,960
Speaker 1: you know, it's all part of business. So you would

53
00:03:02,040 --> 00:03:04,360
Speaker 1: rather have a means to send all this information in

54
00:03:04,360 --> 00:03:07,280
Speaker 1: a way so that even if someone were to intercept

55
00:03:07,320 --> 00:03:12,680
Speaker 1: the messages, no one who didn't you know, didn't have authority,

56
00:03:12,760 --> 00:03:16,280
Speaker 1: would ever understand what it actually says. Now that particular

57
00:03:16,320 --> 00:03:19,880
Speaker 1: idea has been around for as long as we have

58
00:03:20,040 --> 00:03:23,799
Speaker 1: had secrets, How do you communicate a secret to someone

59
00:03:23,840 --> 00:03:27,280
Speaker 1: else without anyone else finding out about it, particularly if

60
00:03:27,320 --> 00:03:30,480
Speaker 1: you can't whisper it quietly to that person. You have

61
00:03:30,600 --> 00:03:33,120
Speaker 1: to come up with a means to hide the meaning

62
00:03:33,360 --> 00:03:35,760
Speaker 1: of your message in some way, and there are a

63
00:03:35,800 --> 00:03:38,640
Speaker 1: lot of different ways to do this. So, for example,

64
00:03:38,680 --> 00:03:41,520
Speaker 1: steganography is a way, and I covered that in a

65
00:03:41,600 --> 00:03:44,000
Speaker 1: previous episode of Tech Stuff with my friend Ariel who

66
00:03:44,080 --> 00:03:48,120
Speaker 1: joined me for that episode. Steganography is the practice of

67
00:03:48,200 --> 00:03:53,040
Speaker 1: hiding a message within some other form of non secret data,

68
00:03:53,400 --> 00:03:58,400
Speaker 1: like uh an image or a music file. You could

69
00:03:58,680 --> 00:04:03,600
Speaker 1: literally high the message inside the image, meaning that unless

70
00:04:03,640 --> 00:04:06,160
Speaker 1: you know what to look for, you're not likely to

71
00:04:06,280 --> 00:04:08,960
Speaker 1: discover the message. That's not the most secure way of

72
00:04:09,000 --> 00:04:11,280
Speaker 1: doing it, obviously, but it is something you can do.

73
00:04:11,400 --> 00:04:15,440
Speaker 1: I remember collecting old Grow the Wanderer comic books by

74
00:04:15,440 --> 00:04:19,920
Speaker 1: Sergio Ariganez, who would hide a secret message in every

75
00:04:19,920 --> 00:04:24,200
Speaker 1: single issue. Typically it would read this is the hidden message.

76
00:04:24,839 --> 00:04:29,760
Speaker 1: So not super duper helpful, but that would be one

77
00:04:29,800 --> 00:04:32,279
Speaker 1: way of doing it. However, you could also hide the

78
00:04:32,279 --> 00:04:37,520
Speaker 1: information within the actual text or or the the the

79
00:04:37,560 --> 00:04:41,080
Speaker 1: file information of the image itself, so instead of like

80
00:04:41,160 --> 00:04:44,200
Speaker 1: hiding it physically inside the image, you're hiding it in

81
00:04:44,240 --> 00:04:47,200
Speaker 1: the code that represents the image. And it's only if

82
00:04:47,240 --> 00:04:49,000
Speaker 1: you know to look at that code that you would

83
00:04:49,000 --> 00:04:51,479
Speaker 1: even see that there was a message there. The whole

84
00:04:51,520 --> 00:04:54,719
Speaker 1: point is you're concealing that secret message inside some other

85
00:04:54,839 --> 00:04:58,400
Speaker 1: non secret information that you can freely distribute, so you

86
00:04:58,440 --> 00:05:01,920
Speaker 1: can send that photo out being reasonably confident that people

87
00:05:01,960 --> 00:05:05,120
Speaker 1: are not going to see that there's a secret message there.

88
00:05:05,160 --> 00:05:06,839
Speaker 1: You only know that it's there if you know to

89
00:05:06,880 --> 00:05:10,919
Speaker 1: look for it. At least that's the theory or the hope. Obviously,

90
00:05:10,960 --> 00:05:14,240
Speaker 1: if someone is super nosy and just curious and they

91
00:05:14,279 --> 00:05:17,520
Speaker 1: start looking at these things more closely, they might uncover

92
00:05:17,640 --> 00:05:22,440
Speaker 1: that message, So it's not always the most secure methodology. Cryptography,

93
00:05:22,480 --> 00:05:27,159
Speaker 1: which means to make communication secret, is a very broad category,

94
00:05:27,360 --> 00:05:31,279
Speaker 1: and encryption falls into that category. Encryption is the specific

95
00:05:31,360 --> 00:05:35,520
Speaker 1: process to make hidden or secret. The two terms are

96
00:05:35,560 --> 00:05:39,440
Speaker 1: often used interchangeably, but technically they are distinct, though I

97
00:05:39,440 --> 00:05:42,159
Speaker 1: guess if enough people use them interchangeably for long enough,

98
00:05:42,560 --> 00:05:45,320
Speaker 1: the meaning itself will change, because that's how language works.

99
00:05:45,360 --> 00:05:49,400
Speaker 1: But that's a bonus episode, I guess. So with encryption,

100
00:05:49,720 --> 00:05:54,480
Speaker 1: you would use some sort of process too encrypt the data.

101
00:05:54,520 --> 00:05:57,720
Speaker 1: You would use a cipher to take a plane message

102
00:05:58,400 --> 00:06:00,960
Speaker 1: the secret that you wish to can municate, and you

103
00:06:00,960 --> 00:06:03,440
Speaker 1: would turn it into something that the average person would

104
00:06:03,440 --> 00:06:07,640
Speaker 1: not be able to understand, but your intended recipient would

105
00:06:07,640 --> 00:06:10,160
Speaker 1: have the knowledge of how to reverse this process to

106
00:06:10,520 --> 00:06:13,560
Speaker 1: decipher the mixed up message, so that he or she

107
00:06:13,680 --> 00:06:17,440
Speaker 1: could read the original secret. So let's take a super

108
00:06:17,480 --> 00:06:21,720
Speaker 1: simple example, one that would never be used in modern

109
00:06:21,760 --> 00:06:26,520
Speaker 1: day encryption. Let's take a basic substitution cipher, in which

110
00:06:26,560 --> 00:06:30,000
Speaker 1: we swap out letters for other letters. In a super

111
00:06:30,040 --> 00:06:33,080
Speaker 1: simple version of this, are two communicators have agreed that,

112
00:06:33,120 --> 00:06:36,360
Speaker 1: for the purposes of their messaging, they will shift all

113
00:06:36,480 --> 00:06:40,640
Speaker 1: letters over by one so that an A will be

114
00:06:40,680 --> 00:06:44,440
Speaker 1: represented by the letter B, A B will be represented

115
00:06:44,440 --> 00:06:47,120
Speaker 1: by the letter C, and so forth until you get

116
00:06:47,160 --> 00:06:49,320
Speaker 1: to Z, which will be represented by the letter A.

117
00:06:49,720 --> 00:06:52,720
Speaker 1: So if I wanted to write out uh j O

118
00:06:52,960 --> 00:06:55,960
Speaker 1: n as my sign off, shortening my name to John,

119
00:06:56,400 --> 00:06:59,279
Speaker 1: I would actually use the letters k P O. My

120
00:06:59,360 --> 00:07:02,400
Speaker 1: recipient would receive this message and say, ah, I must

121
00:07:02,640 --> 00:07:07,919
Speaker 1: decipher it by cleverly stepping back each letter by one position,

122
00:07:08,040 --> 00:07:10,440
Speaker 1: and I get j O N Oh it was Jonathan

123
00:07:10,440 --> 00:07:13,960
Speaker 1: who sent me this message. Now, clearly anyone could crack

124
00:07:14,040 --> 00:07:17,040
Speaker 1: that code. It would not take any sort of master

125
00:07:17,360 --> 00:07:19,440
Speaker 1: code breaker to do it. You don't need a computer

126
00:07:19,520 --> 00:07:24,440
Speaker 1: to do it. Is the simplest of substitution ciphers, and

127
00:07:25,680 --> 00:07:29,720
Speaker 1: obviously that would never stand. However, more much more complicated

128
00:07:29,720 --> 00:07:32,800
Speaker 1: ciphers have been used throughout history, and in fact, one

129
00:07:32,840 --> 00:07:37,600
Speaker 1: of the earliest uses for computers was in decoding encrypted messages.

130
00:07:37,680 --> 00:07:41,200
Speaker 1: During World War Two, computers were dedicated to cracking codes

131
00:07:41,240 --> 00:07:45,160
Speaker 1: made by physical devices like the Enigma machine. But I've

132
00:07:45,160 --> 00:07:47,480
Speaker 1: talked about that in previous episodes, so I'm gonna move

133
00:07:47,520 --> 00:07:51,640
Speaker 1: on rather than dwell on it here now. In cryptology,

134
00:07:51,800 --> 00:07:57,320
Speaker 1: people often use examples to explain specific implementations and strategies.

135
00:07:57,960 --> 00:08:01,480
Speaker 1: This has led to two fictional care is becoming placeholders

136
00:08:01,520 --> 00:08:03,920
Speaker 1: for these examples. So when you want to say person

137
00:08:04,000 --> 00:08:07,760
Speaker 1: A in person B, that gets really cumbersome. So those

138
00:08:07,840 --> 00:08:11,520
Speaker 1: characters are Alice and Bob. If you've ever heard examples

139
00:08:11,640 --> 00:08:16,160
Speaker 1: using Alice and Bob, it comes from this branch of cryptology.

140
00:08:16,280 --> 00:08:20,640
Speaker 1: The characters were actually created by three researchers named Ron Revest,

141
00:08:20,960 --> 00:08:24,760
Speaker 1: Audi Schamir, and Leonard Alderman, who came up with the

142
00:08:24,960 --> 00:08:28,160
Speaker 1: r S A encryption strategy. R s A for the

143
00:08:28,200 --> 00:08:33,400
Speaker 1: first initial uh of each of their last names, so Revest, Schamir,

144
00:08:33,520 --> 00:08:36,160
Speaker 1: and Ottoman you get r S A. And when they

145
00:08:36,160 --> 00:08:38,080
Speaker 1: came up with that, they decided they needed to have

146
00:08:38,280 --> 00:08:42,520
Speaker 1: examples to describe what their whole procedure was, and they

147
00:08:42,800 --> 00:08:45,400
Speaker 1: invented these characters of Alice and Bob. More on R

148
00:08:45,520 --> 00:08:47,680
Speaker 1: s A in just a minute. Alice and Bob have

149
00:08:47,800 --> 00:08:51,560
Speaker 1: become the arc types for cryptological discussions and beyond that. Actually,

150
00:08:51,559 --> 00:08:53,800
Speaker 1: Alie and Bob are used in lots of different examples

151
00:08:53,800 --> 00:08:57,440
Speaker 1: for tech, not just cryptology, but basically, the premise spoils

152
00:08:57,480 --> 00:09:00,319
Speaker 1: down to Bob and Alice both wished to commune dicate

153
00:09:00,360 --> 00:09:03,600
Speaker 1: with one another without other people being able to intercept

154
00:09:03,640 --> 00:09:07,200
Speaker 1: and understand those messages. All right, So Alice and Bob

155
00:09:07,760 --> 00:09:09,960
Speaker 1: they want to send messages to each other using some

156
00:09:10,000 --> 00:09:12,880
Speaker 1: sort of computer device. It could be an actual computer

157
00:09:13,000 --> 00:09:15,360
Speaker 1: like a desktop or a laptop. It could be a smartphone.

158
00:09:15,400 --> 00:09:18,200
Speaker 1: It could be a laptop, tablet, computer, could be any

159
00:09:18,200 --> 00:09:23,320
Speaker 1: of these things. It doesn't really matter that the point

160
00:09:23,400 --> 00:09:25,559
Speaker 1: is that they want to be able to send electronic

161
00:09:25,640 --> 00:09:29,000
Speaker 1: messages in some meaningful way. And when Alice and Bob

162
00:09:29,040 --> 00:09:33,360
Speaker 1: send messages back and forth, their devices aren't doing so directly. Right,

163
00:09:33,400 --> 00:09:38,200
Speaker 1: there's no direct connection between Alice's device and Bob's device

164
00:09:38,280 --> 00:09:41,000
Speaker 1: unless they happen to be very close together and able

165
00:09:41,000 --> 00:09:43,840
Speaker 1: to use some sort of technology like near field communication

166
00:09:43,960 --> 00:09:47,160
Speaker 1: or NFC. Apart from that, where you're so close you

167
00:09:47,160 --> 00:09:49,760
Speaker 1: could just whisper it to each other. Let's say that

168
00:09:49,840 --> 00:09:52,480
Speaker 1: you are across the country from one. Alice is over

169
00:09:52,520 --> 00:09:55,400
Speaker 1: in California, Bob is in New York. The messages they

170
00:09:55,440 --> 00:09:59,080
Speaker 1: send to each other are going to go through routers

171
00:09:59,080 --> 00:10:02,480
Speaker 1: and switches and servers, and eventually they're going to pass

172
00:10:02,520 --> 00:10:06,160
Speaker 1: through some sort of central server system before going through

173
00:10:06,320 --> 00:10:09,320
Speaker 1: even more routers and switches and servers and eventually ending

174
00:10:09,400 --> 00:10:12,760
Speaker 1: up on the other person's phone. So whatever service Alice

175
00:10:12,760 --> 00:10:18,360
Speaker 1: and Bob are using, that centralized server, which is the

176
00:10:18,360 --> 00:10:21,359
Speaker 1: the you know, kind of like the the Great uh

177
00:10:21,480 --> 00:10:25,560
Speaker 1: Post service for that particular app it the message has

178
00:10:25,600 --> 00:10:28,440
Speaker 1: to pass through there because this is a specific app

179
00:10:28,520 --> 00:10:32,679
Speaker 1: being used offered by a specific company. So Alice sends

180
00:10:32,720 --> 00:10:36,960
Speaker 1: Bob a message on their agreed upon messaging app. Alice's

181
00:10:36,960 --> 00:10:39,440
Speaker 1: message is going to head out over the Internet, hit

182
00:10:39,480 --> 00:10:42,080
Speaker 1: the server in charge of handling the messaging service that

183
00:10:42,160 --> 00:10:44,720
Speaker 1: they are both using, and then continue on until it

184
00:10:44,760 --> 00:10:47,360
Speaker 1: hits Bob. Now, if the message is in plain text,

185
00:10:47,559 --> 00:10:52,120
Speaker 1: meaning it's unencrypted, anyone along that pathway could potentially read

186
00:10:52,160 --> 00:10:55,720
Speaker 1: the contents of that message. That includes hackers who could

187
00:10:55,720 --> 00:10:59,360
Speaker 1: have compromised a system somewhere along that pathway, and they're

188
00:10:59,360 --> 00:11:01,800
Speaker 1: trying to pull a man in the middle attack. The message,

189
00:11:01,800 --> 00:11:04,880
Speaker 1: in other words, is not safe. One way to fix

190
00:11:04,920 --> 00:11:08,440
Speaker 1: that is through a simple encryption method in which Alice

191
00:11:08,480 --> 00:11:12,640
Speaker 1: and Bob each have their own private encryption and decryption

192
00:11:12,720 --> 00:11:16,560
Speaker 1: keys with the server. So let's say Alice and Bob

193
00:11:16,840 --> 00:11:19,760
Speaker 1: are using a messaging app. Let's call the app x

194
00:11:19,880 --> 00:11:23,040
Speaker 1: y Z, and x y Z has its server systems.

195
00:11:23,520 --> 00:11:26,440
Speaker 1: When Alice wants to send Bob a message, she uses

196
00:11:26,520 --> 00:11:29,679
Speaker 1: her own personal x y Z encryption key to do

197
00:11:29,760 --> 00:11:33,520
Speaker 1: so and sends it along. No one else other than

198
00:11:33,720 --> 00:11:37,000
Speaker 1: x y Z has that encryption key, so Alice and

199
00:11:37,160 --> 00:11:40,280
Speaker 1: x y Z share it, but nobody else does. Her

200
00:11:40,360 --> 00:11:44,400
Speaker 1: message will get to the server, which sees that Alice

201
00:11:44,440 --> 00:11:47,520
Speaker 1: wants to send a message to Bob. So the service says,

202
00:11:47,559 --> 00:11:50,480
Speaker 1: all right, I'll send this to Bob, except Bob can't

203
00:11:50,520 --> 00:11:54,400
Speaker 1: read it because it's been encrypted with Alice's key, so

204
00:11:54,720 --> 00:11:57,679
Speaker 1: Bob doesn't have Alice's key. So what the server then

205
00:11:57,720 --> 00:12:01,880
Speaker 1: does will decrypt the message because it has Alice's key

206
00:12:01,920 --> 00:12:05,120
Speaker 1: and it also has Bob's key, so it then re

207
00:12:05,360 --> 00:12:09,320
Speaker 1: encrypts the message using Bob's key and then sends that

208
00:12:09,480 --> 00:12:14,640
Speaker 1: along to Bob, so the message has been encrypted twice. Technically,

209
00:12:14,679 --> 00:12:18,640
Speaker 1: it was encrypted, decrypted at the server level, encrypted again

210
00:12:18,800 --> 00:12:22,760
Speaker 1: sent to Bob. Whenever the message is passing over the Internet,

211
00:12:23,200 --> 00:12:27,600
Speaker 1: it is encrypted, which is pretty safe. But you may say, well,

212
00:12:27,640 --> 00:12:31,240
Speaker 1: what about the time that it's spending on the server.

213
00:12:32,280 --> 00:12:35,680
Speaker 1: That's the rub. See, this is not end to end encryption.

214
00:12:35,720 --> 00:12:38,840
Speaker 1: I'll explain that a little bit later. The server is

215
00:12:38,880 --> 00:12:42,559
Speaker 1: able to decrypt all incoming messages, no matter who is

216
00:12:42,600 --> 00:12:45,679
Speaker 1: sending them, and then encrypt them with the respective keys

217
00:12:45,720 --> 00:12:49,240
Speaker 1: belonging to whoever was the intended recipients. But this creates

218
00:12:49,240 --> 00:12:53,880
Speaker 1: a few different problematic scenarios. One is that governments they

219
00:12:54,080 --> 00:12:58,040
Speaker 1: love this strategy because it opens up the possibility that

220
00:12:58,080 --> 00:13:01,440
Speaker 1: the server could release message is on a court order

221
00:13:01,640 --> 00:13:05,600
Speaker 1: or some other means, Like if the government orders the

222
00:13:05,800 --> 00:13:09,600
Speaker 1: service to share those messages and the services compelled to

223
00:13:09,840 --> 00:13:14,920
Speaker 1: agree to it, then potentially the government could get hold

224
00:13:15,040 --> 00:13:18,640
Speaker 1: of unencrypted plane text messages because at the server level

225
00:13:18,679 --> 00:13:21,720
Speaker 1: everything gets unlocked. So the government agency can go to

226
00:13:21,840 --> 00:13:24,559
Speaker 1: X Y Z and say, we suspect Alice and Bob

227
00:13:24,600 --> 00:13:27,520
Speaker 1: are plotting something dangerous. We have some evidence, but we

228
00:13:27,559 --> 00:13:30,280
Speaker 1: want access to the messages they're sending to each other.

229
00:13:30,920 --> 00:13:33,480
Speaker 1: Lives could be at stake, and then they flash a

230
00:13:33,480 --> 00:13:36,480
Speaker 1: Warren or something, and because X y Z is following

231
00:13:36,480 --> 00:13:39,559
Speaker 1: this strategy, it could hand over such information if ordered to,

232
00:13:39,760 --> 00:13:43,400
Speaker 1: revealing all the messages between Alice and Bob. And maybe

233
00:13:43,440 --> 00:13:46,280
Speaker 1: Alice and Bob really were plotting something terrible and it's

234
00:13:46,320 --> 00:13:49,920
Speaker 1: prevented as a result, or maybe they're just happy people

235
00:13:50,080 --> 00:13:53,199
Speaker 1: and their private messages have now been compromised and they

236
00:13:53,200 --> 00:13:57,880
Speaker 1: were completely innocent, and yet there otherwise private communication has

237
00:13:57,920 --> 00:14:01,880
Speaker 1: now been made at least semi public. The strategy also

238
00:14:01,920 --> 00:14:06,040
Speaker 1: creates an incredibly attractive target for hackers, obviously, because if

239
00:14:06,080 --> 00:14:09,400
Speaker 1: you can compromise that central server, you can get at

240
00:14:09,440 --> 00:14:11,880
Speaker 1: all the data that's going across it because it gets

241
00:14:12,000 --> 00:14:15,160
Speaker 1: decrypted there. In fact, this has happened a few times

242
00:14:15,160 --> 00:14:17,480
Speaker 1: in the past where hackers have managed to get access

243
00:14:17,559 --> 00:14:20,880
Speaker 1: to such systems and mind them for data and dump

244
00:14:20,920 --> 00:14:23,520
Speaker 1: all the information onto various places on the web, or

245
00:14:23,880 --> 00:14:26,720
Speaker 1: more likely on the dark web. This is where stuff

246
00:14:26,760 --> 00:14:30,600
Speaker 1: like credit card information or compromising photos can come into play,

247
00:14:30,640 --> 00:14:34,160
Speaker 1: where hackers have managed to get into a server and

248
00:14:34,160 --> 00:14:37,880
Speaker 1: and get all that information. Sure the messages were encrypted

249
00:14:37,960 --> 00:14:40,120
Speaker 1: as they went over the Internet, but that one central

250
00:14:40,160 --> 00:14:44,280
Speaker 1: point everything was unlocked and ready for exploitation and to

251
00:14:44,600 --> 00:14:48,720
Speaker 1: end Encryption aims to defeat this by creating a strategy

252
00:14:48,800 --> 00:14:51,840
Speaker 1: in which only Alice and Bob know how to decrypt

253
00:14:51,920 --> 00:14:55,080
Speaker 1: messages from the other. They have a private means of

254
00:14:55,120 --> 00:14:59,520
Speaker 1: decrypting information. The central server does not have access to

255
00:14:59,600 --> 00:15:03,360
Speaker 1: that it decryption key, and so the messages sent across

256
00:15:03,440 --> 00:15:07,320
Speaker 1: the server are meaningless to the server, is just gobbledygook.

257
00:15:07,800 --> 00:15:11,680
Speaker 1: The server could confirm that messages had passed between Alice

258
00:15:11,720 --> 00:15:13,760
Speaker 1: and Bob, you could say, well, yes, we know that

259
00:15:13,880 --> 00:15:17,640
Speaker 1: the two have communicated with each other, but they the

260
00:15:17,640 --> 00:15:19,840
Speaker 1: company wouldn't be able to speak to what was actually

261
00:15:19,880 --> 00:15:22,960
Speaker 1: in those messages. So if a government agency served up

262
00:15:22,960 --> 00:15:25,480
Speaker 1: a search warrant, all they would get is some garbled,

263
00:15:25,480 --> 00:15:28,920
Speaker 1: seemingly random and meaningless data. They would need that private

264
00:15:29,040 --> 00:15:31,600
Speaker 1: key to make any sense of it. Now, there are

265
00:15:31,600 --> 00:15:35,560
Speaker 1: companies that absolutely love using the strategy because it really

266
00:15:35,600 --> 00:15:37,880
Speaker 1: takes the heat off of them that not only are

267
00:15:37,880 --> 00:15:41,320
Speaker 1: they able to offer their customers a secure way to communicate,

268
00:15:41,840 --> 00:15:43,840
Speaker 1: there's no way for the company to know what sort

269
00:15:43,880 --> 00:15:47,320
Speaker 1: of data is crossing its servers, so it cannot be

270
00:15:47,440 --> 00:15:51,480
Speaker 1: complicit in anything illegal because it doesn't know what's happening.

271
00:15:51,960 --> 00:15:54,680
Speaker 1: It provides a service one in which people may communicate

272
00:15:54,720 --> 00:15:57,040
Speaker 1: with one another in a secure fashion, and that's it.

273
00:15:57,520 --> 00:16:00,320
Speaker 1: But it raises a question how the heck are Alice

274
00:16:00,360 --> 00:16:03,400
Speaker 1: and Bob is supposed to get a private key across

275
00:16:03,440 --> 00:16:06,320
Speaker 1: to each other? Right? If Alice and Bob are apart

276
00:16:06,400 --> 00:16:10,040
Speaker 1: from one another and their communication requires going through a

277
00:16:10,080 --> 00:16:13,360
Speaker 1: centralized service, how do they establish a secret means of

278
00:16:13,360 --> 00:16:17,400
Speaker 1: communication in the first place. If that first message is hey,

279
00:16:17,600 --> 00:16:21,640
Speaker 1: let's use this secret code, but it's decrypted by the

280
00:16:21,640 --> 00:16:25,360
Speaker 1: centralized server, then the centralized server knows the secret code,

281
00:16:25,400 --> 00:16:27,760
Speaker 1: it doesn't help. Well. The answer comes down to what

282
00:16:27,880 --> 00:16:34,080
Speaker 1: was called an asymmetric cryptographic algorithm that uses public key cryptography.

283
00:16:34,120 --> 00:16:37,000
Speaker 1: And in this method, there are two keys that are

284
00:16:37,160 --> 00:16:40,920
Speaker 1: used for each message. One of them is a public key,

285
00:16:40,960 --> 00:16:44,040
Speaker 1: which is unique to each user that can be freely

286
00:16:44,280 --> 00:16:47,840
Speaker 1: distributed across the internet. The other is a private key,

287
00:16:47,880 --> 00:16:51,640
Speaker 1: which is a jealously guarded secret that allows only one

288
00:16:51,800 --> 00:16:55,240
Speaker 1: entity the chance to decrypt information. I'll speak about more

289
00:16:55,240 --> 00:16:57,880
Speaker 1: of this in just a minute, but first let's take

290
00:16:57,880 --> 00:17:07,840
Speaker 1: a quick break to thank our sponsor. Alright, So how

291
00:17:07,880 --> 00:17:12,160
Speaker 1: does public key cryptography work. Each person using the system,

292
00:17:12,240 --> 00:17:15,240
Speaker 1: let's call it ABC. For this case, we're talking about

293
00:17:15,240 --> 00:17:19,080
Speaker 1: different service from x y Z. Every single person using

294
00:17:19,119 --> 00:17:24,199
Speaker 1: ABC gets too encryption keys, or well, one encryption key

295
00:17:24,240 --> 00:17:28,440
Speaker 1: and one decryption key. Technically, one key is the person's

296
00:17:28,480 --> 00:17:31,639
Speaker 1: private key, So Alice has a private key that's unique

297
00:17:31,680 --> 00:17:34,280
Speaker 1: to her, and Bob has a private key it's unique

298
00:17:34,320 --> 00:17:38,439
Speaker 1: to him. They they also each have a public key,

299
00:17:38,640 --> 00:17:42,360
Speaker 1: so Alice has a public key, Bob has another public key.

300
00:17:42,440 --> 00:17:45,400
Speaker 1: These are also unique to Alice and Bob, but those

301
00:17:45,480 --> 00:17:49,040
Speaker 1: keys are published publicly on the service across the Internet.

302
00:17:49,440 --> 00:17:52,600
Speaker 1: Anyone who wants to send an encrypted message to Alice

303
00:17:52,720 --> 00:17:56,560
Speaker 1: or Bob must use their respective public keys. In other words,

304
00:17:56,560 --> 00:17:58,359
Speaker 1: if I want to send Alice a message, I have

305
00:17:58,400 --> 00:18:02,960
Speaker 1: to use Alice's public to encrypt it. To decode a message,

306
00:18:03,320 --> 00:18:06,240
Speaker 1: you have to have the private key. So if you

307
00:18:06,280 --> 00:18:09,240
Speaker 1: know both keys for a specific message, you can transform

308
00:18:09,240 --> 00:18:12,320
Speaker 1: that garbled mess into meaningful communication. And because you keep

309
00:18:12,359 --> 00:18:15,520
Speaker 1: your own private key to yourself, at least ideally, the

310
00:18:15,560 --> 00:18:18,399
Speaker 1: only person who can decode messages meant for you is you.

311
00:18:19,119 --> 00:18:23,280
Speaker 1: So Alice takes Bob's public key and uses it to

312
00:18:23,400 --> 00:18:26,960
Speaker 1: encode her message to Bob. She then sends the encoded

313
00:18:27,000 --> 00:18:31,080
Speaker 1: message to Bob across the messaging service. The encoded message

314
00:18:31,080 --> 00:18:35,080
Speaker 1: passes through ABC's server, which cannot read the message because

315
00:18:35,119 --> 00:18:39,080
Speaker 1: ABC does not have Bob's private key. Only Bob has

316
00:18:39,119 --> 00:18:42,520
Speaker 1: that the message gets to Bob, he uses his private

317
00:18:42,600 --> 00:18:45,400
Speaker 1: key and it decodes the message and then he can

318
00:18:45,440 --> 00:18:48,040
Speaker 1: read it in plain text. But what is actually going

319
00:18:48,080 --> 00:18:50,439
Speaker 1: on here? How is this happening? Well that that depends

320
00:18:50,480 --> 00:18:54,480
Speaker 1: heavily on the specific cryptographic strategy used, But we can

321
00:18:54,600 --> 00:18:58,399
Speaker 1: go with one of the earliest proposed methods of doing this,

322
00:18:58,560 --> 00:19:01,600
Speaker 1: the RSA algorithm, the one I mentioned earlier that was

323
00:19:01,760 --> 00:19:04,200
Speaker 1: proposed by the guys who used Alice and Bob as

324
00:19:04,200 --> 00:19:08,320
Speaker 1: examples in the first place. This one relies on prime numbers,

325
00:19:08,760 --> 00:19:12,200
Speaker 1: and just a quick refresher, a prime number is only

326
00:19:12,280 --> 00:19:15,320
Speaker 1: equal to one times the prime number itself. There are

327
00:19:15,320 --> 00:19:18,480
Speaker 1: no other multiple multiple factors. There are no other numbers

328
00:19:18,520 --> 00:19:21,119
Speaker 1: you can multiply together to get the prime number. So

329
00:19:21,160 --> 00:19:23,119
Speaker 1: if you can get to the value of a number

330
00:19:23,240 --> 00:19:26,439
Speaker 1: by by multiplying any other two numbers, it's what we

331
00:19:26,520 --> 00:19:30,080
Speaker 1: call a composite number. And is not a prime number.

332
00:19:30,359 --> 00:19:34,840
Speaker 1: So the numbers one, two, and three are all prime numbers.

333
00:19:34,880 --> 00:19:37,399
Speaker 1: They are only divisible by themselves and by one that

334
00:19:37,520 --> 00:19:41,560
Speaker 1: there are no other factors, right That our whole numbers anyway,

335
00:19:41,840 --> 00:19:44,119
Speaker 1: because the only multiplication you can do to arrive at

336
00:19:44,160 --> 00:19:46,720
Speaker 1: them is one times one is one, one times two

337
00:19:46,760 --> 00:19:49,879
Speaker 1: is two, and one times three is three, respectively. Four

338
00:19:50,080 --> 00:19:53,680
Speaker 1: is the smallest composite number because you can multiply two

339
00:19:53,720 --> 00:19:56,600
Speaker 1: times two to get four. So it's one times four

340
00:19:56,680 --> 00:19:58,360
Speaker 1: is four and two times two is four. That means

341
00:19:58,400 --> 00:20:00,680
Speaker 1: it is not a prime number. It's a posit number.

342
00:20:00,840 --> 00:20:03,400
Speaker 1: So the number is only divisible by one or itself,

343
00:20:03,600 --> 00:20:06,359
Speaker 1: then it's a prime number. The r s a algorithm

344
00:20:06,400 --> 00:20:12,520
Speaker 1: takes two really really really big prime numbers, like hundreds

345
00:20:12,520 --> 00:20:14,679
Speaker 1: of digits long. You might find systems that use a

346
00:20:14,680 --> 00:20:17,760
Speaker 1: five twelve bit prime number, though a lot of security

347
00:20:17,800 --> 00:20:20,359
Speaker 1: experts would say a thousand twenty four bit prime number

348
00:20:20,359 --> 00:20:24,919
Speaker 1: would be better than It takes those two already huge numbers. Remember,

349
00:20:24,960 --> 00:20:28,720
Speaker 1: it has to identify two different prime numbers that are

350
00:20:28,920 --> 00:20:33,959
Speaker 1: enormous and then multiplies those two enormous prime numbers together

351
00:20:34,240 --> 00:20:38,000
Speaker 1: to get a new number. This number is called the modulus.

352
00:20:38,440 --> 00:20:42,240
Speaker 1: It then requires the calculation of the totient of that product.

353
00:20:43,160 --> 00:20:45,080
Speaker 1: And this is where you say, wait a minute, hang on,

354
00:20:45,200 --> 00:20:47,199
Speaker 1: what the heck is a totient? At least if you

355
00:20:47,240 --> 00:20:49,600
Speaker 1: are like me, I was an English lit major, I

356
00:20:49,680 --> 00:20:52,720
Speaker 1: never got to totients in my studies in math. But

357
00:20:52,840 --> 00:20:57,520
Speaker 1: a totient is a number um that describes the number

358
00:20:57,520 --> 00:21:01,919
Speaker 1: of integers smaller than a given number that are co

359
00:21:02,040 --> 00:21:04,280
Speaker 1: prime with that number. Oh, that's a lot of numbers.

360
00:21:04,359 --> 00:21:06,119
Speaker 1: Let me try that again. Let's say you've got a

361
00:21:06,160 --> 00:21:12,160
Speaker 1: really big number, and the totient is all of the

362
00:21:12,200 --> 00:21:15,720
Speaker 1: integers that are smaller than this really big number that

363
00:21:15,760 --> 00:21:19,560
Speaker 1: are co prime with that really big number, which means

364
00:21:19,600 --> 00:21:23,639
Speaker 1: they share no factors except one with that really big number.

365
00:21:24,359 --> 00:21:28,119
Speaker 1: So let's use an example, because this sounds really really vague. Right.

366
00:21:28,240 --> 00:21:31,439
Speaker 1: Let's say that we have the number fourteen. We we

367
00:21:31,560 --> 00:21:35,280
Speaker 1: got to fourteen, uh, and we took prime numbers. We

368
00:21:35,320 --> 00:21:37,680
Speaker 1: took the number two and the number seven and we

369
00:21:37,800 --> 00:21:40,480
Speaker 1: multiplied those together. Those are our prime numbers we started with.

370
00:21:40,560 --> 00:21:42,880
Speaker 1: So again, this is just an example. You would never

371
00:21:43,000 --> 00:21:46,080
Speaker 1: use numbers this small. You multiply two times seven, you

372
00:21:46,080 --> 00:21:49,520
Speaker 1: get fourteen. Fourteen is your modulus. Then you say what

373
00:21:49,680 --> 00:21:54,399
Speaker 1: is the totient of fourteen? The answer is six, and

374
00:21:54,480 --> 00:21:58,919
Speaker 1: here is why. First, you take all the integers that

375
00:21:59,000 --> 00:22:02,960
Speaker 1: are less than four team, which is one through thirteen. Right,

376
00:22:03,240 --> 00:22:05,560
Speaker 1: you can't use fourteen. You have to use all the

377
00:22:05,600 --> 00:22:07,480
Speaker 1: ones that are all the whole numbers that are lower

378
00:22:07,480 --> 00:22:10,640
Speaker 1: than fourteen, so one through thirteen. But the numbers also

379
00:22:10,680 --> 00:22:13,280
Speaker 1: have to be co prime with fourteen. That means they

380
00:22:13,400 --> 00:22:17,880
Speaker 1: cannot share any factors that fourteen has. Now, the factors

381
00:22:17,920 --> 00:22:20,280
Speaker 1: of fourteen are two and seven, so we get those

382
00:22:20,359 --> 00:22:23,159
Speaker 1: off the list those those get removed, So two and

383
00:22:23,200 --> 00:22:25,600
Speaker 1: seven are gone. But you also have to eliminate the

384
00:22:25,680 --> 00:22:30,440
Speaker 1: numbers four, six, eight, ten, and twelve. All of those

385
00:22:30,520 --> 00:22:34,200
Speaker 1: numbers have two as a factor, and fourteen also has

386
00:22:34,200 --> 00:22:37,159
Speaker 1: two as a factor, so those numbers are not co prime,

387
00:22:37,600 --> 00:22:40,280
Speaker 1: so you strike those. Once you've gotten rid of those,

388
00:22:40,600 --> 00:22:43,400
Speaker 1: the numbers you have left that are lower than fourteen

389
00:22:43,480 --> 00:22:49,760
Speaker 1: are one, three, five, nine, eleven, and thirteen. Now nine

390
00:22:49,840 --> 00:22:52,160
Speaker 1: is not a prime number, right, You can multiply three

391
00:22:52,160 --> 00:22:55,760
Speaker 1: times three and get nine, but it is co prime

392
00:22:56,119 --> 00:23:00,080
Speaker 1: with fourteen because the two numbers share no common factor

393
00:23:00,240 --> 00:23:03,120
Speaker 1: except for the number one. So when you count up

394
00:23:03,280 --> 00:23:06,439
Speaker 1: all the co primes that are left. After you've eliminated

395
00:23:06,520 --> 00:23:09,080
Speaker 1: the numbers that are not co primes, you find out

396
00:23:09,160 --> 00:23:14,760
Speaker 1: you've got six numbers total one, three, five, nteen. That's

397
00:23:14,840 --> 00:23:18,960
Speaker 1: six numbers. So the totient for fourteen is six. By

398
00:23:18,960 --> 00:23:21,320
Speaker 1: the way, there's actually a formula for calculating the totent

399
00:23:21,400 --> 00:23:24,600
Speaker 1: of your huge number that's really really easy. And what

400
00:23:24,720 --> 00:23:27,239
Speaker 1: you do is you take your first prime number, the

401
00:23:27,240 --> 00:23:29,720
Speaker 1: one that you want. You know, in our case we

402
00:23:29,800 --> 00:23:31,680
Speaker 1: use two and seven. Let's say you take your first

403
00:23:31,680 --> 00:23:35,720
Speaker 1: prime number, you subtract one from that. You take your

404
00:23:35,760 --> 00:23:39,600
Speaker 1: second big prime number, you subtract one from that, and

405
00:23:39,640 --> 00:23:42,480
Speaker 1: you multiply those two new numbers together, and that is

406
00:23:42,520 --> 00:23:45,080
Speaker 1: your totent. So in our example, we take our two

407
00:23:45,080 --> 00:23:48,240
Speaker 1: prime numbers of two and seven, we subtract one from each.

408
00:23:48,400 --> 00:23:50,600
Speaker 1: That means we have a one and a six. We

409
00:23:50,720 --> 00:23:54,199
Speaker 1: multiply these two new numbers together one time, six is six.

410
00:23:54,760 --> 00:23:57,399
Speaker 1: That's the totent for fourteen. It's the fastest way to

411
00:23:57,440 --> 00:24:01,159
Speaker 1: calculate the totent. Next, because we're not done yet, you

412
00:24:01,240 --> 00:24:04,160
Speaker 1: have to select an integer, but not just any integer.

413
00:24:04,720 --> 00:24:07,240
Speaker 1: The integer you select has to be larger than the

414
00:24:07,320 --> 00:24:11,480
Speaker 1: number one, but smaller than the totient of your big

415
00:24:11,480 --> 00:24:14,199
Speaker 1: old number that you generated earlier by multiplying those two

416
00:24:14,240 --> 00:24:16,719
Speaker 1: prime numbers together and then taking the totient from it.

417
00:24:17,320 --> 00:24:19,719
Speaker 1: In fact, it has to be a co prime with

418
00:24:19,760 --> 00:24:24,119
Speaker 1: the totient and the modulus itself, so it cannot share

419
00:24:24,359 --> 00:24:27,560
Speaker 1: a factor with the totient or with the modulus. So again,

420
00:24:27,600 --> 00:24:30,679
Speaker 1: in my example, where I had fourteen as our modulus,

421
00:24:30,920 --> 00:24:34,320
Speaker 1: the totient was six, we need an integer between one

422
00:24:34,359 --> 00:24:38,240
Speaker 1: and six that is co prime with both six and fourteen.

423
00:24:38,760 --> 00:24:40,960
Speaker 1: So we cannot use one because the number has to

424
00:24:40,960 --> 00:24:44,240
Speaker 1: be bigger than one. We can't use two or three

425
00:24:44,600 --> 00:24:49,240
Speaker 1: or four because all of those share factors with six. Right,

426
00:24:49,359 --> 00:24:52,359
Speaker 1: it has to be co prime, so you cannot use those.

427
00:24:52,760 --> 00:24:56,240
Speaker 1: The only number available to us is five. This becomes

428
00:24:56,280 --> 00:24:59,719
Speaker 1: our encryption key. Figuring out the decryption key is a

429
00:24:59,720 --> 00:25:02,920
Speaker 1: little more complicated, and that's probably making some of you

430
00:25:03,200 --> 00:25:05,280
Speaker 1: scratch your heads, because what I just went through is

431
00:25:05,480 --> 00:25:07,520
Speaker 1: fairly complicated for those of us who are not very

432
00:25:07,560 --> 00:25:10,959
Speaker 1: mathematically inclined. I include myself in that number. By the way,

433
00:25:11,240 --> 00:25:14,040
Speaker 1: the prime number we arrived at for the encryption key

434
00:25:14,160 --> 00:25:17,159
Speaker 1: has the greatest common divisor or g c D of

435
00:25:17,280 --> 00:25:20,600
Speaker 1: one with the totient of that number. So we take

436
00:25:20,640 --> 00:25:24,480
Speaker 1: the multiplicative inverse of this number with respect to the

437
00:25:24,480 --> 00:25:28,200
Speaker 1: totient using the extended Euclidean algorithm. Yeah, I get super

438
00:25:28,280 --> 00:25:30,760
Speaker 1: duper complicated, and I'm pretty sure I can't explain it,

439
00:25:31,440 --> 00:25:35,919
Speaker 1: especially not in audio alone, so the well will skip ahead.

440
00:25:36,240 --> 00:25:39,680
Speaker 1: You just imagine you do some more complicated math and

441
00:25:39,760 --> 00:25:42,479
Speaker 1: you denote this result with the letter D, and that

442
00:25:42,560 --> 00:25:46,560
Speaker 1: represents the private key. Now. I use the prime numbers

443
00:25:46,560 --> 00:25:48,520
Speaker 1: two and seven in my example only because they were

444
00:25:48,520 --> 00:25:51,000
Speaker 1: easy to work with, But that's precisely why no one

445
00:25:51,040 --> 00:25:53,840
Speaker 1: in the right mind would ever use those to encrypt anything.

446
00:25:54,320 --> 00:25:56,320
Speaker 1: You want to have a pretty large number to work with,

447
00:25:56,520 --> 00:25:59,000
Speaker 1: and even your public key should be pretty big. A

448
00:25:59,119 --> 00:26:03,480
Speaker 1: standard public key is sixty five thousand, five seven, which

449
00:26:03,520 --> 00:26:05,520
Speaker 1: is a prime number, and as long as it has

450
00:26:05,560 --> 00:26:07,800
Speaker 1: no common factors with the modulus, you're good to go.

451
00:26:08,040 --> 00:26:10,120
Speaker 1: And the only way it could have a common factor

452
00:26:10,160 --> 00:26:13,479
Speaker 1: with a modulus is if it was itself a factor

453
00:26:13,520 --> 00:26:15,919
Speaker 1: of your big number, but that's not likely to happen.

454
00:26:16,280 --> 00:26:18,000
Speaker 1: So you want a big key, but you don't want

455
00:26:18,040 --> 00:26:19,840
Speaker 1: it to be too big, and the reason for that

456
00:26:20,000 --> 00:26:22,840
Speaker 1: is encryption is more efficient if you're using smaller numbers

457
00:26:22,840 --> 00:26:26,880
Speaker 1: as your keys. Remember, like essentially encryption involves doing lots

458
00:26:26,880 --> 00:26:29,280
Speaker 1: of math problems. The bigger the numbers are for your

459
00:26:29,320 --> 00:26:31,880
Speaker 1: math problems, the bigger the results are going to be,

460
00:26:32,040 --> 00:26:34,119
Speaker 1: and the larger it's going to make your message. So

461
00:26:34,160 --> 00:26:37,199
Speaker 1: if you're trying to encrypt a very long message to someone,

462
00:26:37,760 --> 00:26:40,920
Speaker 1: then ends up being a much larger amount of data,

463
00:26:41,320 --> 00:26:45,640
Speaker 1: like sometimes several times larger than the original amount of data.

464
00:26:45,680 --> 00:26:47,560
Speaker 1: And if you get too big, you may not even

465
00:26:47,600 --> 00:26:49,920
Speaker 1: be able to send it if the service you're using

466
00:26:49,960 --> 00:26:53,920
Speaker 1: has a limit on the size of messages. So if

467
00:26:53,960 --> 00:26:58,120
Speaker 1: you go too big, it gets inefficient and clunky and

468
00:26:58,200 --> 00:27:00,439
Speaker 1: sometimes too big for you to even hand. But if

469
00:27:00,440 --> 00:27:02,720
Speaker 1: you go too small, you risk the possibility of someone

470
00:27:02,760 --> 00:27:05,600
Speaker 1: being able to suss out the private key given enough

471
00:27:05,600 --> 00:27:08,360
Speaker 1: time and processing powers, So it's a balancing act. One

472
00:27:08,400 --> 00:27:10,679
Speaker 1: thing you could do to make things move more quickly

473
00:27:11,040 --> 00:27:13,520
Speaker 1: is to use this public and private key approach to

474
00:27:13,680 --> 00:27:17,679
Speaker 1: establish a new encryption key between the two communicators. This

475
00:27:17,720 --> 00:27:21,000
Speaker 1: would be called a session key. So Alice would send

476
00:27:21,000 --> 00:27:24,560
Speaker 1: Bob a message, she would use Bob's public encryption key

477
00:27:24,600 --> 00:27:27,359
Speaker 1: that's and she would encrypt a message that says, hey,

478
00:27:27,440 --> 00:27:29,960
Speaker 1: got a second and she sends it to Bob, and

479
00:27:30,040 --> 00:27:32,840
Speaker 1: Bob receives the message. He uses his private key, he

480
00:27:33,000 --> 00:27:35,840
Speaker 1: decodes the message and sends a message back to Alice

481
00:27:35,960 --> 00:27:39,040
Speaker 1: using her public encryption key that says, sure I do.

482
00:27:39,560 --> 00:27:42,679
Speaker 1: And here's a brand new private key that we can

483
00:27:42,760 --> 00:27:46,119
Speaker 1: use for each other for the purposes of this conversation session.

484
00:27:46,680 --> 00:27:49,640
Speaker 1: And then Alice would decrypt this message, and they would

485
00:27:49,720 --> 00:27:53,520
Speaker 1: each have a private key that they could use to

486
00:27:53,600 --> 00:27:57,400
Speaker 1: each other, a symmetric key. All subsequent messages between Bob

487
00:27:57,400 --> 00:28:01,320
Speaker 1: and Alice would use this new private encryption methodology. And

488
00:28:01,359 --> 00:28:04,800
Speaker 1: we call this symmetric because you're using the same key

489
00:28:04,840 --> 00:28:09,119
Speaker 1: to encrypt and decrypt and UH. Otherwise, you wouldn't be

490
00:28:09,119 --> 00:28:12,120
Speaker 1: able to send this to each other because it would

491
00:28:12,160 --> 00:28:14,520
Speaker 1: be public information and everyone would have a copy of

492
00:28:14,560 --> 00:28:18,199
Speaker 1: the key. Using the public private key message to first

493
00:28:18,800 --> 00:28:22,240
Speaker 1: send this is clever because the first message, anyone could

494
00:28:22,280 --> 00:28:24,720
Speaker 1: intercept it, but they're not gonna know what's in it.

495
00:28:25,200 --> 00:28:28,120
Speaker 1: And then all subsequent messages would be using a totally

496
00:28:28,200 --> 00:28:34,439
Speaker 1: different UH encryption methodology. So even if someone were to

497
00:28:34,640 --> 00:28:39,239
Speaker 1: monitor this communication, the communications they would see would be

498
00:28:39,360 --> 00:28:43,760
Speaker 1: encrypted in different ways, and that would make it practically

499
00:28:43,800 --> 00:28:46,280
Speaker 1: impossible to figure out what was going on. You could

500
00:28:46,280 --> 00:28:49,320
Speaker 1: even go a step further and have a new key

501
00:28:49,360 --> 00:28:52,920
Speaker 1: generated with essentially every single message, so it's essentially a

502
00:28:52,920 --> 00:28:56,160
Speaker 1: one ahead. Each message gets a new key to be

503
00:28:56,280 --> 00:29:00,280
Speaker 1: encrypted by, and since it keeps changing each time someone

504
00:29:00,280 --> 00:29:03,240
Speaker 1: sends a message, like Bob sends Alice a new message

505
00:29:03,240 --> 00:29:08,160
Speaker 1: and says uh uh secret, secret, secret, brand new key,

506
00:29:08,200 --> 00:29:11,200
Speaker 1: Alice uses the brand new key to part her message

507
00:29:11,200 --> 00:29:13,520
Speaker 1: in too, Bob, and it says more secrets, more secrets,

508
00:29:13,520 --> 00:29:18,160
Speaker 1: more secrets, new, brand new key. Bob uses the uh

509
00:29:18,240 --> 00:29:22,320
Speaker 1: the the key from one session earlier to unlock Alice's

510
00:29:22,560 --> 00:29:26,120
Speaker 1: message reads it then uses the brand new key to

511
00:29:26,280 --> 00:29:28,800
Speaker 1: send the next message. You could keep doing this forever.

512
00:29:29,320 --> 00:29:31,960
Speaker 1: You could keep sending a brand new key with every

513
00:29:32,000 --> 00:29:35,040
Speaker 1: single message, and it would make your communications very secure.

514
00:29:35,080 --> 00:29:37,400
Speaker 1: It would be a little slow because you would have

515
00:29:37,480 --> 00:29:41,640
Speaker 1: to do this decryption encryption thing every single step, and

516
00:29:41,760 --> 00:29:44,040
Speaker 1: you wouldn't be able to repeat the steps because they'll

517
00:29:44,080 --> 00:29:46,640
Speaker 1: be using a new key every time, but it would

518
00:29:46,640 --> 00:29:49,160
Speaker 1: be really secure, and in fact, there are some messaging

519
00:29:49,200 --> 00:29:53,520
Speaker 1: apps that use this kind of methodology. So the really

520
00:29:53,520 --> 00:29:57,120
Speaker 1: big benefit of symmetric encryption is that. Well, one, it's

521
00:29:57,200 --> 00:30:00,520
Speaker 1: faster than using public and private key approach, but another

522
00:30:00,640 --> 00:30:03,000
Speaker 1: is that limits the number of times that you use

523
00:30:03,080 --> 00:30:06,520
Speaker 1: the public key, and that can be a good thing. Now,

524
00:30:06,520 --> 00:30:10,160
Speaker 1: it's not easy with today's computers, but at least there

525
00:30:10,240 --> 00:30:14,040
Speaker 1: it's at least theoretically possible that hackers could suss out

526
00:30:14,240 --> 00:30:16,960
Speaker 1: a public encryption key and work backward to figure out

527
00:30:16,960 --> 00:30:20,440
Speaker 1: the private encryption key. It's a non trivial problem. It

528
00:30:20,480 --> 00:30:23,640
Speaker 1: requires a ton of computer processing power, and it's not

529
00:30:23,720 --> 00:30:26,800
Speaker 1: likely to happen if you're using really secure encryption. But

530
00:30:26,880 --> 00:30:30,120
Speaker 1: the more frequently use the same public key, the more

531
00:30:30,200 --> 00:30:32,120
Speaker 1: data hackers have to work with, and they can start

532
00:30:32,200 --> 00:30:38,440
Speaker 1: looking for patterns. Patterns are the bane of encryption. When

533
00:30:38,440 --> 00:30:41,920
Speaker 1: you have patterns, then you can start to establish rules.

534
00:30:41,960 --> 00:30:43,800
Speaker 1: And when you start to establish rules, you can start

535
00:30:43,840 --> 00:30:46,600
Speaker 1: working backward and figuring out what generates those rules, and

536
00:30:46,600 --> 00:30:50,680
Speaker 1: eventually you figure out the methodology for encoding the information.

537
00:30:51,440 --> 00:30:55,640
Speaker 1: So you don't want to have patterns in your encoded information.

538
00:30:55,720 --> 00:30:59,280
Speaker 1: If possible, if you keep using the public private key

539
00:30:59,320 --> 00:31:02,959
Speaker 1: method and it's all you're using, then hackers eventually can

540
00:31:03,000 --> 00:31:06,560
Speaker 1: gather enough information that if they have a sufficiently powerful

541
00:31:06,600 --> 00:31:11,800
Speaker 1: computer and enough time on their hands, they can decrypt it. Uh.

542
00:31:11,840 --> 00:31:14,160
Speaker 1: That if is a big one though, because if you're

543
00:31:14,280 --> 00:31:20,000
Speaker 1: using pretty strong encryption, it would take weeks or months

544
00:31:20,200 --> 00:31:24,400
Speaker 1: or more likely years to decrypt the messages. Now, in

545
00:31:24,440 --> 00:31:28,800
Speaker 1: a recent episode, I talked about quantum computers. With a

546
00:31:28,840 --> 00:31:31,760
Speaker 1: classical computer, like I said, figuring out prime numbers for

547
00:31:31,800 --> 00:31:35,160
Speaker 1: a really large number takes an incredibly long time. Depending

548
00:31:35,160 --> 00:31:37,160
Speaker 1: on the size of the number, it could take like

549
00:31:37,200 --> 00:31:39,840
Speaker 1: I said, years for a classical computer to sort it out.

550
00:31:39,880 --> 00:31:43,760
Speaker 1: But a quantum computer, if it's sufficiently powerful, could solve

551
00:31:43,760 --> 00:31:49,200
Speaker 1: those problems much more quickly using a process called Shore's algorithm.

552
00:31:49,560 --> 00:31:51,400
Speaker 1: So if you have a quantum computer with a sufficient

553
00:31:51,480 --> 00:31:54,680
Speaker 1: number of cubits, you could crack this type of encryption

554
00:31:54,720 --> 00:31:58,480
Speaker 1: relatively quickly and pretty reliably. So it's therefore imperative to

555
00:31:58,520 --> 00:32:01,040
Speaker 1: look at new methods of encrypt in order to avoid

556
00:32:01,080 --> 00:32:04,200
Speaker 1: a situation in which the first really powerful quantum computer

557
00:32:04,400 --> 00:32:08,200
Speaker 1: effectively gets a skeleton key to all encrypted messages that

558
00:32:08,240 --> 00:32:11,080
Speaker 1: have been sent across the Internet. Now, let's talk a

559
00:32:11,120 --> 00:32:14,360
Speaker 1: second about the history of end to end encryption, and

560
00:32:14,400 --> 00:32:16,520
Speaker 1: then a bit more about the apps and services that

561
00:32:16,720 --> 00:32:23,720
Speaker 1: use it and some that do not use it. In nine, Whitfield,

562
00:32:23,720 --> 00:32:27,120
Speaker 1: Diffy and Martin Hellman proposed the concept of using public

563
00:32:27,160 --> 00:32:32,080
Speaker 1: and private key combinations in order to distribute symmetric encryption keys.

564
00:32:32,480 --> 00:32:35,240
Speaker 1: It's possible they weren't the first to consider this too.

565
00:32:35,400 --> 00:32:38,560
Speaker 1: There's been some evidence to suggest the British Secret Service

566
00:32:38,600 --> 00:32:40,960
Speaker 1: had come up with a similar approach but never really

567
00:32:41,000 --> 00:32:43,719
Speaker 1: did anything with it. Their idea was the basis not

568
00:32:43,800 --> 00:32:46,480
Speaker 1: just for the r s A algorithm, but a few

569
00:32:46,480 --> 00:32:49,720
Speaker 1: others as well, like the el Camal crypto system, which

570
00:32:49,760 --> 00:32:52,800
Speaker 1: was named after Tahir el Camal, or the d s

571
00:32:52,840 --> 00:32:56,960
Speaker 1: A system also known as the Digital Signature algorithm, invented

572
00:32:57,000 --> 00:33:01,360
Speaker 1: by a guy named David Kravitz, as well as the

573
00:33:01,440 --> 00:33:05,520
Speaker 1: Diffie Hellman cryptosystem, which obviously was created by Whitfield, Diffie

574
00:33:05,520 --> 00:33:09,680
Speaker 1: and Martin Hellman. Then along came p g P, which

575
00:33:09,720 --> 00:33:14,560
Speaker 1: stands for Pretty Good Privacy. The cryptosystem is a hybrid system.

576
00:33:14,720 --> 00:33:20,200
Speaker 1: It was proposed by Phil Zimmerman in n Zimmerman graduated

577
00:33:20,280 --> 00:33:24,080
Speaker 1: from Florida Atlantic University with a degree in computer science,

578
00:33:24,480 --> 00:33:27,480
Speaker 1: and he was active in a project called the Freeze,

579
00:33:27,600 --> 00:33:31,760
Speaker 1: also known as the Nuclear Weapons Freeze campaign. The purpose

580
00:33:31,880 --> 00:33:35,080
Speaker 1: of this organization was to try and curtail nuclear arms

581
00:33:35,120 --> 00:33:39,000
Speaker 1: production in an effort to de escalate mounting global tensions

582
00:33:39,120 --> 00:33:43,120
Speaker 1: and to remove a A an existential threat to the

583
00:33:43,200 --> 00:33:47,280
Speaker 1: human race, and Zimmerman created p g P to allow

584
00:33:47,320 --> 00:33:50,239
Speaker 1: for secure email communications among various parties to make their

585
00:33:50,280 --> 00:33:54,040
Speaker 1: efforts more effective and less likely to get snooped on. Now,

586
00:33:54,040 --> 00:33:56,880
Speaker 1: when you use p g P, the first thing it

587
00:33:56,920 --> 00:34:00,760
Speaker 1: does is it takes your plain text message and impresses it,

588
00:34:01,440 --> 00:34:03,800
Speaker 1: and so it makes it smaller, which is one benefit,

589
00:34:03,880 --> 00:34:07,440
Speaker 1: but it also helps to make it more secure because

590
00:34:07,440 --> 00:34:10,440
Speaker 1: it reduces patterns that might otherwise be useful to hackers

591
00:34:10,480 --> 00:34:14,879
Speaker 1: who want to decrypt messages. It also generates a session key,

592
00:34:14,920 --> 00:34:17,640
Speaker 1: that's that symmetric encryption key I was just talking about

593
00:34:17,680 --> 00:34:19,880
Speaker 1: that would be used throughout the length of any particular

594
00:34:19,920 --> 00:34:23,239
Speaker 1: communication session. The way PGP does this is pretty cool.

595
00:34:23,280 --> 00:34:26,720
Speaker 1: It actually generates the data for the session key based

596
00:34:26,719 --> 00:34:29,600
Speaker 1: on your mouse movements and key strokes you've typed, so

597
00:34:29,719 --> 00:34:32,520
Speaker 1: it's based upon physical interactions you've had with your computer,

598
00:34:33,200 --> 00:34:37,520
Speaker 1: not with any other just random prime number. All of this,

599
00:34:37,719 --> 00:34:40,600
Speaker 1: the compressed plane text message and the unique session key

600
00:34:40,640 --> 00:34:43,959
Speaker 1: generated from your physical interactions with your computer then get

601
00:34:44,120 --> 00:34:49,440
Speaker 1: encoded using this public private key strategy. If Alice sends

602
00:34:49,440 --> 00:34:53,520
Speaker 1: a message using PGP too Bob, first, Alice's computer will

603
00:34:53,520 --> 00:34:57,200
Speaker 1: compress your message, it will append a session key based

604
00:34:57,200 --> 00:35:00,919
Speaker 1: on Alice's typing and mouse movements, and then encrypt all

605
00:35:01,000 --> 00:35:04,640
Speaker 1: of that mess using Bob's public key. Then the message

606
00:35:04,760 --> 00:35:08,280
Speaker 1: travels to Bob. Bob uses his private key to decode

607
00:35:08,280 --> 00:35:11,320
Speaker 1: the message and he can see the session key Alice

608
00:35:11,320 --> 00:35:13,719
Speaker 1: has set up and then use that to communicate back

609
00:35:13,760 --> 00:35:16,880
Speaker 1: with her securely for the rest of the session. Zimmerman's

610
00:35:16,920 --> 00:35:20,040
Speaker 1: p GP was more than just pretty good. It was

611
00:35:20,120 --> 00:35:23,359
Speaker 1: actually a brilliant approach to encryption. It became adopted by

612
00:35:23,360 --> 00:35:26,640
Speaker 1: many organizations and people across the United States and then

613
00:35:26,680 --> 00:35:29,040
Speaker 1: the world, and that caused a huge amount of trouble

614
00:35:29,160 --> 00:35:32,760
Speaker 1: for Zimmerman. For one thing, r s A Security wished

615
00:35:32,760 --> 00:35:35,960
Speaker 1: to question Zimmerman about the use of the r s

616
00:35:36,040 --> 00:35:38,799
Speaker 1: A algorithm within p GP. There was some disputes about

617
00:35:38,840 --> 00:35:43,080
Speaker 1: the licensing. Then the United States Custom Services decided to

618
00:35:43,080 --> 00:35:46,840
Speaker 1: investigate Zimmerman because of the distribution of p GP beyond

619
00:35:46,920 --> 00:35:51,160
Speaker 1: the US borders. Because in the United States there was

620
00:35:51,239 --> 00:35:56,120
Speaker 1: the Arms Control Act and it listed cryptographics software as

621
00:35:56,160 --> 00:35:59,600
Speaker 1: a type of munitions, which would mean if Zimmerman had

622
00:35:59,600 --> 00:36:02,080
Speaker 1: allowed his work to go outside the US, he would

623
00:36:02,080 --> 00:36:04,840
Speaker 1: be in very big trouble because it was similar to

624
00:36:04,960 --> 00:36:09,000
Speaker 1: shipping prohibited weapons outside the country. And if you think

625
00:36:09,000 --> 00:36:11,920
Speaker 1: that sounds a little crazy that software could be considered

626
00:36:11,920 --> 00:36:15,960
Speaker 1: a weapon, Welcome to the twenty first century. Zimmerman was

627
00:36:16,160 --> 00:36:19,640
Speaker 1: never officially charged with any crime, but the investigation did

628
00:36:19,800 --> 00:36:23,359
Speaker 1: last several years. Eventually, the courts determined that p GP

629
00:36:23,560 --> 00:36:26,680
Speaker 1: did not fall into a category that would be considered munitions,

630
00:36:26,960 --> 00:36:30,440
Speaker 1: and today his PGP technology is the property of the

631
00:36:30,480 --> 00:36:34,560
Speaker 1: security firm Symantec, which purchased the PGP Corporation back in

632
00:36:35,520 --> 00:36:36,719
Speaker 1: Now I've got a little bit more I want to

633
00:36:36,719 --> 00:36:39,200
Speaker 1: say about end to end encryption, but before I get

634
00:36:39,239 --> 00:36:42,080
Speaker 1: to that, let's take another quick break to thank our sponsor.

635
00:36:49,200 --> 00:36:53,080
Speaker 1: So the PGP strategy also allows for digital signatures, which

636
00:36:53,080 --> 00:36:54,920
Speaker 1: are a way for you to make sure the message

637
00:36:54,960 --> 00:36:58,319
Speaker 1: being sent to you hasn't been intercepted and altered in

638
00:36:58,360 --> 00:37:00,840
Speaker 1: any way. Plus you can be sure that the message

639
00:37:00,920 --> 00:37:03,960
Speaker 1: came from the person it claims to have come from.

640
00:37:04,120 --> 00:37:07,000
Speaker 1: Digital signatures and p g P use what is called

641
00:37:07,040 --> 00:37:11,120
Speaker 1: a one way hash function. The implementation p g P

642
00:37:11,360 --> 00:37:13,640
Speaker 1: uses is to take the message, which can be of

643
00:37:13,680 --> 00:37:17,640
Speaker 1: any length. So let's say it's a really really long message.

644
00:37:17,880 --> 00:37:20,840
Speaker 1: Then it does a mathematical process on it to arrive

645
00:37:20,920 --> 00:37:24,800
Speaker 1: at a fixed length output, such as one sixty bits.

646
00:37:24,840 --> 00:37:28,080
Speaker 1: So the one sixty bits does not contain the entire message.

647
00:37:28,080 --> 00:37:32,200
Speaker 1: It represents the entire message. It's fine distinction. It doesn't

648
00:37:32,239 --> 00:37:34,960
Speaker 1: matter how long that original message was. The goal is

649
00:37:35,000 --> 00:37:39,160
Speaker 1: to create this shorter hash that speeds things up. It

650
00:37:39,239 --> 00:37:43,000
Speaker 1: doesn't make file sizes balloon from the encryption process. The

651
00:37:43,040 --> 00:37:47,439
Speaker 1: hash function depends entirely on the message itself, So if

652
00:37:47,480 --> 00:37:50,680
Speaker 1: someone were to change even one single bit as in one,

653
00:37:50,800 --> 00:37:54,280
Speaker 1: zero or one of information in that message, the hash

654
00:37:54,400 --> 00:37:58,240
Speaker 1: value would also change because it depends upon the nature

655
00:37:58,280 --> 00:38:00,479
Speaker 1: of the rest of the message. So if you send

656
00:38:00,480 --> 00:38:02,960
Speaker 1: a message to someone and they know what the hash

657
00:38:03,040 --> 00:38:05,960
Speaker 1: value is supposed to be, they can verify that the

658
00:38:06,000 --> 00:38:09,960
Speaker 1: associated message was never altered. So they've got essentially the

659
00:38:09,960 --> 00:38:12,640
Speaker 1: the hash value it's supposed to be and the hash

660
00:38:12,719 --> 00:38:15,360
Speaker 1: value it actually is. If those two numbers are different,

661
00:38:15,760 --> 00:38:18,480
Speaker 1: then they know that something has happened to the message

662
00:38:18,640 --> 00:38:22,000
Speaker 1: in transit. The way PGP does this is to generate

663
00:38:22,040 --> 00:38:25,720
Speaker 1: a hash called the message digest. It uses the message

664
00:38:25,840 --> 00:38:29,279
Speaker 1: digest and the private key to create a signature. Then

665
00:38:29,360 --> 00:38:32,160
Speaker 1: p GP sends the plain text in signature together to

666
00:38:32,320 --> 00:38:36,120
Speaker 1: the recipient. The recipient uses PGP to recompute the message

667
00:38:36,200 --> 00:38:39,080
Speaker 1: digest to verify it is from the supposed center and

668
00:38:39,080 --> 00:38:42,400
Speaker 1: then it hasn't been altered. And if the recomputed message

669
00:38:42,480 --> 00:38:45,200
Speaker 1: digest is the same as the original that was in

670
00:38:45,239 --> 00:38:49,160
Speaker 1: the message, everything's good. If the recomputed one is not

671
00:38:49,280 --> 00:38:52,399
Speaker 1: the same, then something has happened, and you know that

672
00:38:52,440 --> 00:38:56,319
Speaker 1: the the trail between the two has been compromised in

673
00:38:56,400 --> 00:38:59,239
Speaker 1: some way. By the way, this does not reveal the

674
00:38:59,280 --> 00:39:02,680
Speaker 1: private key to the other person. The private key remains

675
00:39:02,719 --> 00:39:05,600
Speaker 1: private um P. G P is able to protect that,

676
00:39:05,680 --> 00:39:09,279
Speaker 1: which is kind of cool. Alright. So who uses end

677
00:39:09,280 --> 00:39:12,320
Speaker 1: to end encryption and who does not? Well? Among the

678
00:39:12,360 --> 00:39:15,680
Speaker 1: messaging apps that use end to end encryption, there is

679
00:39:15,800 --> 00:39:19,880
Speaker 1: the Big Daddy What's App. That's an incredibly popular messaging app.

680
00:39:19,920 --> 00:39:23,440
Speaker 1: It is owned by Facebook. Now what didn't start off

681
00:39:23,480 --> 00:39:26,879
Speaker 1: as a Facebook project. It was purchased by Facebook. The

682
00:39:26,920 --> 00:39:31,799
Speaker 1: original messaging app was co founded by Jan Kom who

683
00:39:31,800 --> 00:39:34,520
Speaker 1: grew up in the Ukraine. And wanted to create a

684
00:39:34,520 --> 00:39:38,080
Speaker 1: way in which people could send messages safely without the

685
00:39:38,120 --> 00:39:41,920
Speaker 1: fear of an oppressive presence, like a totalitarian government agency,

686
00:39:41,920 --> 00:39:44,840
Speaker 1: for example, that could read all of them. Now, he

687
00:39:44,880 --> 00:39:47,919
Speaker 1: had moved to America when he was sixteen, and years

688
00:39:48,000 --> 00:39:50,560
Speaker 1: later he met with a guy named Brian Acton and

689
00:39:50,640 --> 00:39:54,360
Speaker 1: Acting became the other co founder of WhatsApp. They created

690
00:39:54,400 --> 00:39:57,200
Speaker 1: the messaging service, and they used end to end encryption

691
00:39:57,280 --> 00:40:02,640
Speaker 1: to protect users privacy. When Faceboo Public purchased WhatsApp rather,

692
00:40:03,320 --> 00:40:07,279
Speaker 1: many security experts express concern because they were worried that

693
00:40:07,320 --> 00:40:11,360
Speaker 1: Facebook might chip away at the messaging apps security and

694
00:40:11,400 --> 00:40:14,480
Speaker 1: allow Facebook the chance to mind these messages and then

695
00:40:14,520 --> 00:40:18,920
Speaker 1: sell that data to advertisers, and there have been accusations

696
00:40:19,040 --> 00:40:23,200
Speaker 1: levied at Facebook claiming as much. To Bias Bolter, a

697
00:40:23,320 --> 00:40:26,920
Speaker 1: security researcher, told reporters at The Guardian that Facebook had

698
00:40:26,960 --> 00:40:30,520
Speaker 1: the ability to generate new encryption keys for offline users

699
00:40:30,600 --> 00:40:34,000
Speaker 1: and thus access messages that would be sent to them.

700
00:40:34,040 --> 00:40:37,080
Speaker 1: Whether that's true or not, I don't know. Facebook has

701
00:40:37,120 --> 00:40:39,760
Speaker 1: denied that they have the ability to read any messages

702
00:40:39,760 --> 00:40:43,120
Speaker 1: from their users, but there's still a lot of concern

703
00:40:43,160 --> 00:40:47,000
Speaker 1: about WhatsApp in general. Another app that uses end to

704
00:40:47,120 --> 00:40:50,800
Speaker 1: end encryption is Signal. Signal started out as text secure

705
00:40:50,880 --> 00:40:53,399
Speaker 1: private messenger, and you can also use Signal to make

706
00:40:53,520 --> 00:40:57,280
Speaker 1: encrypted voice and video calls to other users. These follow

707
00:40:57,360 --> 00:41:01,680
Speaker 1: a similar approach to messages and email els um. They

708
00:41:01,719 --> 00:41:05,840
Speaker 1: protocols are obviously a little different because they're different media

709
00:41:06,040 --> 00:41:11,360
Speaker 1: or different forms of messaging, but the general principle remains

710
00:41:11,400 --> 00:41:14,560
Speaker 1: the same, the idea that all the information being sent

711
00:41:14,719 --> 00:41:17,840
Speaker 1: across these channels is encrypted all the way until it

712
00:41:17,840 --> 00:41:22,560
Speaker 1: gets to the destination device. Whire is a another popular app,

713
00:41:22,680 --> 00:41:26,720
Speaker 1: particularly in countries belonging to the European Union. It's also

714
00:41:26,880 --> 00:41:30,080
Speaker 1: free and open source. The app does store a record

715
00:41:30,120 --> 00:41:33,200
Speaker 1: of the people you've communicated with. That's a little problematic

716
00:41:33,239 --> 00:41:37,120
Speaker 1: from a security standpoint. Now, the purpose of storing this

717
00:41:37,239 --> 00:41:40,960
Speaker 1: information is to make it easier to synchronize and experience

718
00:41:41,000 --> 00:41:43,920
Speaker 1: across devices, because one of the big drawbacks of the

719
00:41:43,960 --> 00:41:47,680
Speaker 1: public private key approach is that ideally you have that

720
00:41:47,719 --> 00:41:51,920
Speaker 1: private key on one and only one device. If you

721
00:41:52,000 --> 00:41:54,840
Speaker 1: share a private key across multiple devices, then you have

722
00:41:55,000 --> 00:41:58,720
Speaker 1: increased the possibility of someone getting access to your private key.

723
00:41:58,760 --> 00:42:01,000
Speaker 1: But if you lose the device pace that the private

724
00:42:01,080 --> 00:42:04,120
Speaker 1: key is on, you lose all your messaging history as well,

725
00:42:04,160 --> 00:42:07,760
Speaker 1: because remember, no one else can see what that encrypted

726
00:42:07,800 --> 00:42:11,000
Speaker 1: stuff is because it all used your public key that

727
00:42:11,080 --> 00:42:13,440
Speaker 1: could only be decoded by your private key. So if

728
00:42:13,480 --> 00:42:16,560
Speaker 1: you lose your private key, all you have is gobbledygook.

729
00:42:16,880 --> 00:42:20,480
Speaker 1: You can't read it. That is, of course, if if

730
00:42:20,520 --> 00:42:24,480
Speaker 1: you haven't been storing messages in some other fashion, uh

731
00:42:24,760 --> 00:42:28,759
Speaker 1: like in plain text some services, you could, as a

732
00:42:28,880 --> 00:42:33,319
Speaker 1: user opt to store your messages using a third party application.

733
00:42:33,760 --> 00:42:37,440
Speaker 1: This would be like you receive the message, you decrypt

734
00:42:37,440 --> 00:42:39,720
Speaker 1: the message when you receive it, you have the message

735
00:42:39,760 --> 00:42:41,799
Speaker 1: in plain text because you have to in order to

736
00:42:41,800 --> 00:42:44,239
Speaker 1: read it or or view it or whatever it may be.

737
00:42:45,040 --> 00:42:47,239
Speaker 1: And then you decide you want to save that message. Well,

738
00:42:47,239 --> 00:42:50,120
Speaker 1: when you're saving that message, you may be saving that

739
00:42:50,280 --> 00:42:53,319
Speaker 1: in plain text using a third party app, in which

740
00:42:53,320 --> 00:42:57,359
Speaker 1: case all that secrecy for the communication is undone by

741
00:42:57,360 --> 00:43:01,200
Speaker 1: your storage. It becomes a security vulnerabile. So you could

742
00:43:01,200 --> 00:43:03,799
Speaker 1: be super careful about messages as they speed to and

743
00:43:03,840 --> 00:43:06,040
Speaker 1: from your device, But if you're then storing everything on

744
00:43:06,080 --> 00:43:09,360
Speaker 1: a cloud server somewhere, you could still technically be allowing

745
00:43:09,400 --> 00:43:12,320
Speaker 1: some other company access to that information. After the fact,

746
00:43:12,840 --> 00:43:15,520
Speaker 1: but if you don't do that, if you lose your phone,

747
00:43:15,560 --> 00:43:18,440
Speaker 1: then you could lose all those messages, all the histories,

748
00:43:18,520 --> 00:43:22,520
Speaker 1: all the communications, So it's a tough situation. Also, by

749
00:43:22,520 --> 00:43:27,319
Speaker 1: the way, this means that the hackers will sometimes concentrate

750
00:43:27,440 --> 00:43:29,800
Speaker 1: not on trying to intercept a message in the middle,

751
00:43:30,160 --> 00:43:33,040
Speaker 1: because if it's encrypted, it could be more trouble than

752
00:43:33,040 --> 00:43:36,480
Speaker 1: what it's worth to try and decrypt it. Or instead

753
00:43:36,480 --> 00:43:40,080
Speaker 1: of doing that, they'll they'll try and target the end device.

754
00:43:40,239 --> 00:43:42,280
Speaker 1: So in other words, they're actually trying to get physical

755
00:43:42,360 --> 00:43:48,080
Speaker 1: access to your computer or phone or somehow be able

756
00:43:48,120 --> 00:43:52,400
Speaker 1: to look at your screen remotely once the decryption process

757
00:43:52,440 --> 00:43:55,040
Speaker 1: has happened, So they want to compromise your device at

758
00:43:55,080 --> 00:43:59,840
Speaker 1: that point because it's relatively easier than trying to do

759
00:44:00,000 --> 00:44:03,000
Speaker 1: crypt one of these heavily encrypted messages on the Internet. Anyway,

760
00:44:03,360 --> 00:44:07,719
Speaker 1: Wire has said that a user once a user deactivates

761
00:44:07,760 --> 00:44:11,759
Speaker 1: deletes his or her account, then the service deletes all

762
00:44:11,880 --> 00:44:16,200
Speaker 1: the associated information with that account, all the stored connections,

763
00:44:16,280 --> 00:44:19,440
Speaker 1: everything is gone once you delete your account with Wire,

764
00:44:19,600 --> 00:44:23,080
Speaker 1: according to the service, there's also a multi platform app

765
00:44:23,160 --> 00:44:27,440
Speaker 1: that's popular that's called Telegram. Telegram, you can search for

766
00:44:27,520 --> 00:44:30,319
Speaker 1: people by user name rather than by phone number. A

767
00:44:30,360 --> 00:44:32,319
Speaker 1: lot of apps require that you know the phone number

768
00:44:32,320 --> 00:44:35,399
Speaker 1: of the person you're trying to contact, but Telegram, when

769
00:44:35,440 --> 00:44:41,600
Speaker 1: you create a profile, in addition to creating a profile

770
00:44:41,640 --> 00:44:44,359
Speaker 1: based on a telephone number, you can create a user name,

771
00:44:44,400 --> 00:44:45,960
Speaker 1: and so people can search for you on that. It

772
00:44:45,960 --> 00:44:49,120
Speaker 1: makes it a little easier to communicate with people. There's

773
00:44:49,160 --> 00:44:52,600
Speaker 1: also another app called wicker w i c k R,

774
00:44:52,800 --> 00:44:57,080
Speaker 1: which is really popular in enterprises, so in business, not

775
00:44:57,160 --> 00:45:01,640
Speaker 1: necessarily as much with individual users. But as for email,

776
00:45:01,680 --> 00:45:07,600
Speaker 1: their services like tour Guard, Hushmail, Mail, fence to to,

777
00:45:07,719 --> 00:45:10,960
Speaker 1: Not to run Box, proton Mail, all of these use

778
00:45:11,040 --> 00:45:13,680
Speaker 1: various encryption strategies to keep messages safe. Not all of

779
00:45:13,680 --> 00:45:17,200
Speaker 1: them necessarily use end to end encryption, some of them

780
00:45:17,280 --> 00:45:19,719
Speaker 1: use other methods for encryption than in too end, but

781
00:45:19,760 --> 00:45:22,560
Speaker 1: they all do some form of encryption on their messages,

782
00:45:23,280 --> 00:45:26,839
Speaker 1: whereas popular email services like Gmail do not support end

783
00:45:26,840 --> 00:45:30,840
Speaker 1: to end encryption, and one reason they might not I

784
00:45:30,880 --> 00:45:33,239
Speaker 1: can't say this is for sure sure, but one reason

785
00:45:33,320 --> 00:45:37,200
Speaker 1: they may not do that is that it's possible that

786
00:45:37,440 --> 00:45:41,160
Speaker 1: they want to mine all that information for the possibility

787
00:45:41,160 --> 00:45:43,879
Speaker 1: of advertising against it. So in other words, you send

788
00:45:43,880 --> 00:45:47,400
Speaker 1: a message. Alison's a message to Bob, Alice talks about

789
00:45:48,480 --> 00:45:51,200
Speaker 1: all these nice flowers that she saw on a recent trip.

790
00:45:51,680 --> 00:45:55,560
Speaker 1: And then Alice starts seeing ads mysteriously whenever she's using

791
00:45:55,560 --> 00:46:00,440
Speaker 1: a Google service that is talking about flowers. Well, that

792
00:46:00,560 --> 00:46:04,319
Speaker 1: raises some very tricky ethical questions. Now, whether or not

793
00:46:04,400 --> 00:46:09,160
Speaker 1: that's actually going on, that's a that's a matter of

794
00:46:09,640 --> 00:46:12,960
Speaker 1: investigation on a case by case basis. You know, not

795
00:46:13,080 --> 00:46:16,880
Speaker 1: all emails are necessarily being mined for information, but they

796
00:46:16,920 --> 00:46:19,680
Speaker 1: all could be if they're not being encrypted end to end.

797
00:46:20,320 --> 00:46:24,600
Speaker 1: So there are a lot of advocates for privacy out

798
00:46:24,600 --> 00:46:28,279
Speaker 1: there who some of them will say, Hey, don't even

799
00:46:28,320 --> 00:46:31,399
Speaker 1: bother sending me any email if it's not encrypted. I'm

800
00:46:31,440 --> 00:46:34,560
Speaker 1: not interested you. If you're not using an encryption service

801
00:46:34,600 --> 00:46:39,200
Speaker 1: of some sort, then I do not want to communicate

802
00:46:39,239 --> 00:46:42,560
Speaker 1: with you because I prefer to keep my privacy private.

803
00:46:43,000 --> 00:46:45,560
Speaker 1: And that's a legitimate argument, I think. I don't think

804
00:46:45,560 --> 00:46:48,840
Speaker 1: it's not necessarily the indication that someone is up to something,

805
00:46:49,560 --> 00:46:53,760
Speaker 1: you know, suspicious, But it's important to remember that without

806
00:46:53,800 --> 00:46:56,400
Speaker 1: that end to end encryption, it is a possibility that

807
00:46:56,520 --> 00:46:59,080
Speaker 1: someone along the way is reading your messages. Besides the

808
00:46:59,080 --> 00:47:01,839
Speaker 1: person you're sending them too. I would like to thank

809
00:47:01,880 --> 00:47:07,520
Speaker 1: our anonymous listener for asking for us to cover this topic.

810
00:47:07,560 --> 00:47:11,080
Speaker 1: It is very fascinating. In the UK in particular, this

811
00:47:11,160 --> 00:47:14,640
Speaker 1: is an ongoing issue where the UK government would love

812
00:47:14,680 --> 00:47:18,520
Speaker 1: to see end to end encryption go away because they

813
00:47:18,719 --> 00:47:23,000
Speaker 1: worry that it's a methodology of UH, that that terrorists

814
00:47:23,040 --> 00:47:25,719
Speaker 1: are using to communicate with one another, and that this

815
00:47:25,840 --> 00:47:29,360
Speaker 1: creates a problem because you can't really investigate those terrorists

816
00:47:29,360 --> 00:47:32,719
Speaker 1: cells at least not through the means of their communication,

817
00:47:32,800 --> 00:47:36,480
Speaker 1: because you can't decrypt it. There have been people who

818
00:47:36,480 --> 00:47:40,160
Speaker 1: have called for a back door of some sort to

819
00:47:40,800 --> 00:47:46,000
Speaker 1: end to end encryption services, which defeats the purpose. Backdoor

820
00:47:46,120 --> 00:47:50,319
Speaker 1: is essentially a vulnerability you introduce on purpose so that

821
00:47:50,600 --> 00:47:55,800
Speaker 1: a party apart from the intended recipient can decode a message,

822
00:47:56,600 --> 00:48:01,680
Speaker 1: and that that raises enormous security problems. Like you don't

823
00:48:01,680 --> 00:48:06,000
Speaker 1: create a backdoor because that means you're introducing the possibility

824
00:48:06,120 --> 00:48:10,640
Speaker 1: of a hacker getting access to this information. You don't

825
00:48:10,680 --> 00:48:15,040
Speaker 1: want to do that. That's that's bad security. So there,

826
00:48:15,400 --> 00:48:19,200
Speaker 1: that's not a great UH strategy moving forward. Is trying

827
00:48:19,200 --> 00:48:22,200
Speaker 1: to force services to create a backdoor. It pretty much

828
00:48:22,520 --> 00:48:26,920
Speaker 1: breaks the whole service in the first place. So I

829
00:48:26,960 --> 00:48:29,560
Speaker 1: don't know what the solution here is. I don't think

830
00:48:29,600 --> 00:48:33,320
Speaker 1: that that getting rid of a technology because some people

831
00:48:33,560 --> 00:48:38,719
Speaker 1: use it improperly is necessarily going to work. Um that

832
00:48:38,840 --> 00:48:42,480
Speaker 1: is uh, it is, it is. It raises a lot

833
00:48:42,480 --> 00:48:44,880
Speaker 1: of questions, and it also means that you have to

834
00:48:44,920 --> 00:48:50,359
Speaker 1: put an awful lot of trust in the governing agency,

835
00:48:50,600 --> 00:48:54,120
Speaker 1: and that can be difficult to do. And also, any

836
00:48:54,360 --> 00:48:57,040
Speaker 1: time you have a back door, anytime you have any

837
00:48:57,239 --> 00:49:01,160
Speaker 1: sort of vulnerability, if people know it exists, then you've

838
00:49:01,200 --> 00:49:04,560
Speaker 1: just given hackers a target to shoot for. And even

839
00:49:04,760 --> 00:49:07,680
Speaker 1: if it works as it's supposed to, where everyone is

840
00:49:07,719 --> 00:49:11,400
Speaker 1: behaving themselves from the good guys standpoint, let's say that

841
00:49:11,440 --> 00:49:13,399
Speaker 1: the government is a good guy in this. In this

842
00:49:13,600 --> 00:49:17,840
Speaker 1: scenario that both Alice and Bob are behaving themselves, the

843
00:49:17,880 --> 00:49:21,239
Speaker 1: government is behaving itself. It's not investigating Alison Bob. It

844
00:49:21,280 --> 00:49:23,759
Speaker 1: has no reason to suspect them. So they're able to

845
00:49:23,800 --> 00:49:28,520
Speaker 1: send their messages securely. But there is this back door

846
00:49:28,640 --> 00:49:33,239
Speaker 1: option that is available, then hackers could target that and

847
00:49:33,280 --> 00:49:35,880
Speaker 1: then they compromise the system, and then Alice and Bob's

848
00:49:35,920 --> 00:49:39,840
Speaker 1: messages are laid bare for the hackers because you've just

849
00:49:39,880 --> 00:49:43,280
Speaker 1: painted a huge target on the service. It's not great,

850
00:49:43,480 --> 00:49:48,880
Speaker 1: and not every end to end encryption service is infallible. Uh.

851
00:49:48,920 --> 00:49:52,400
Speaker 1: There have been security experts who have accused Apple in

852
00:49:52,440 --> 00:49:57,200
Speaker 1: particular of implementing a poor end to end encryption strategy

853
00:49:57,320 --> 00:50:00,879
Speaker 1: for their Eye Message app and have stated that there

854
00:50:00,880 --> 00:50:06,200
Speaker 1: are ways that that could be defeated using classical computer systems.

855
00:50:07,320 --> 00:50:10,000
Speaker 1: So it is possible to do this in a way

856
00:50:10,000 --> 00:50:14,799
Speaker 1: that isn't effective, but the actual idea itself is incredibly

857
00:50:15,000 --> 00:50:19,280
Speaker 1: effective if you implement it properly and you're using classical

858
00:50:19,320 --> 00:50:24,839
Speaker 1: computers as your potential threat. Once quantum computers actually become

859
00:50:24,840 --> 00:50:27,319
Speaker 1: on part of the scene for real z s, when

860
00:50:27,360 --> 00:50:31,000
Speaker 1: they're sufficiently powerful, it'll be a totally different story because

861
00:50:32,200 --> 00:50:38,160
Speaker 1: a sufficiently powerful quantum computer will completely lay bare all

862
00:50:38,280 --> 00:50:40,480
Speaker 1: these encryption schemes, So you have to come up with

863
00:50:40,520 --> 00:50:42,440
Speaker 1: something new. But I talked a little bit about that

864
00:50:42,920 --> 00:50:45,160
Speaker 1: in one of the IBM Think episodes, so you can

865
00:50:45,160 --> 00:50:46,440
Speaker 1: go back and listen to that if you want to

866
00:50:46,520 --> 00:50:49,319
Speaker 1: learn more. As for you, guys, if you have any

867
00:50:49,320 --> 00:50:52,440
Speaker 1: suggestions for future episodes of Tech Stuff, I I welcome

868
00:50:52,480 --> 00:50:56,120
Speaker 1: you to send those suggestions to me. You can email

869
00:50:56,160 --> 00:50:59,120
Speaker 1: the show that addresses tech stuff at how stuff Works

870
00:50:59,160 --> 00:51:01,719
Speaker 1: dot com, or you can drop me a line on

871
00:51:01,760 --> 00:51:06,120
Speaker 1: Twitter or Facebook. The handle at both of those is

872
00:51:06,200 --> 00:51:09,840
Speaker 1: tech stuff hs W. Remember to follow us on Instagram,

873
00:51:09,920 --> 00:51:14,000
Speaker 1: and remember typically I record these episodes live on twitch

874
00:51:14,080 --> 00:51:17,680
Speaker 1: dot tv slash tech stuff on Wednesdays and Fridays. This

875
00:51:17,719 --> 00:51:20,480
Speaker 1: particular episode I'm recording right now was an exception because

876
00:51:20,560 --> 00:51:23,919
Speaker 1: again I'm in a hotel, a music hotel WiFi, which

877
00:51:23,960 --> 00:51:26,719
Speaker 1: is not the greatest thing to use for streaming, but

878
00:51:27,239 --> 00:51:30,600
Speaker 1: usually I stream live from the studio on Wednesdays and Fridays.

879
00:51:30,640 --> 00:51:32,839
Speaker 1: Go to Twitch dot tv slash tech Stuff. You'll see

880
00:51:32,840 --> 00:51:34,360
Speaker 1: the schedule there. I hope to see you in the

881
00:51:34,440 --> 00:51:37,560
Speaker 1: chat room and you can chat with me as I'm recording.

882
00:51:37,880 --> 00:51:39,520
Speaker 1: You can have a grand old time and I'll talk

883
00:51:39,560 --> 00:51:48,440
Speaker 1: to you again really soon. For more on this and

884
00:51:48,480 --> 00:52:01,040
Speaker 1: bathands of other topics. Is that how stuff Works dot com.