1
00:00:00,880 --> 00:00:04,960
Speaker 1: You can always tell when something's been written by AI with.

2
00:00:05,120 --> 00:00:08,480
Speaker 2: Little human involvement. You know what I mean?

3
00:00:09,039 --> 00:00:12,879
Speaker 1: A LinkedIn post that starts with in today's fast paced world,

4
00:00:13,560 --> 00:00:16,560
Speaker 1: or an article that ends with it's not about X,

5
00:00:16,720 --> 00:00:20,800
Speaker 1: it's about why. A paragraph but technically fine, but somehow

6
00:00:20,920 --> 00:00:25,120
Speaker 1: feels like it was written by nobody. And that is

7
00:00:25,200 --> 00:00:29,960
Speaker 1: the uncanny valley of AI writing. And the problem isn't

8
00:00:30,000 --> 00:00:33,839
Speaker 1: just that it sounds robotic, it's that people stop reading.

9
00:00:34,800 --> 00:00:38,519
Speaker 1: The frustrating part is that this happens even when you

10
00:00:38,600 --> 00:00:42,559
Speaker 1: have done the thinking, You write the ideas, hand it

11
00:00:42,560 --> 00:00:46,080
Speaker 1: to AI for a cleanup, and it AI effies the

12
00:00:46,200 --> 00:00:50,199
Speaker 1: whole thing, flattening your voice and coding it with M

13
00:00:50,280 --> 00:00:54,840
Speaker 1: dashes until nothing sounds like you anymore. So today Neo

14
00:00:55,000 --> 00:00:58,560
Speaker 1: and I get into exactly how to stop that from happening.

15
00:00:59,360 --> 00:01:02,800
Speaker 1: We cover specific instructions you can give to AI to

16
00:01:02,920 --> 00:01:07,400
Speaker 1: keep your writings sounding human, things like mixing sentence lengths,

17
00:01:07,520 --> 00:01:10,200
Speaker 1: avoiding stiff transitions, and banning the.

18
00:01:10,200 --> 00:01:12,640
Speaker 2: Rule of three. To see what I did there, I

19
00:01:12,680 --> 00:01:15,400
Speaker 2: just choose it. We also go through.

20
00:01:15,319 --> 00:01:18,560
Speaker 1: Some of the most overused AI words and phrases to

21
00:01:18,680 --> 00:01:23,360
Speaker 1: strike from your prompts entirely and now as a little bonus,

22
00:01:23,440 --> 00:01:25,720
Speaker 1: you will find a link in the show notes to

23
00:01:25,800 --> 00:01:28,800
Speaker 1: our anti AI slot prompt, which is a ready to

24
00:01:28,880 --> 00:01:31,760
Speaker 1: use resource you can customize with your own personal pet

25
00:01:31,800 --> 00:01:36,000
Speaker 1: hates and run any content through before it goes out. Okay,

26
00:01:36,160 --> 00:01:45,640
Speaker 1: let's get on with the show. Welcome to how IAI

27
00:01:45,880 --> 00:01:49,440
Speaker 1: with me, Doctor Amantha Imber and Neo Applin, head of

28
00:01:49,600 --> 00:01:53,640
Speaker 1: Inventium AI. Each episode we share one practical way to

29
00:01:53,760 --> 00:01:55,200
Speaker 1: use AI better at.

30
00:01:55,080 --> 00:01:56,440
Speaker 2: Work and in life.

31
00:01:56,840 --> 00:01:59,920
Speaker 1: No fluff, no tech jargon, just things you can use

32
00:02:00,000 --> 00:02:05,000
Speaker 1: straight away. So much content in the world is clearly

33
00:02:05,000 --> 00:02:07,360
Speaker 1: written by AI now, and I know that it.

34
00:02:07,800 --> 00:02:10,560
Speaker 2: Really frustrates us, But like, why is it a problem?

35
00:02:10,680 --> 00:02:13,560
Speaker 1: Why should we actually be aiming to not sound like

36
00:02:13,680 --> 00:02:16,760
Speaker 1: AI when we are all using AI to create the

37
00:02:16,840 --> 00:02:19,720
Speaker 1: content and words we are putting out into the world.

38
00:02:19,840 --> 00:02:22,400
Speaker 3: Now, Yeah, I think it's really two parts. One is

39
00:02:22,480 --> 00:02:24,720
Speaker 3: don't be AI, in other words, don't just get AI

40
00:02:24,760 --> 00:02:26,799
Speaker 3: to do all your work for you. And then the

41
00:02:26,840 --> 00:02:29,160
Speaker 3: second part is don't sound like AI. And I think

42
00:02:29,200 --> 00:02:31,679
Speaker 3: because a lot of people are now doing things where

43
00:02:31,680 --> 00:02:33,560
Speaker 3: they're going like five dot points and say write me

44
00:02:33,600 --> 00:02:35,280
Speaker 3: a big article, and it might be there are five

45
00:02:35,320 --> 00:02:37,919
Speaker 3: pages that really Ai wrote all of it apart from

46
00:02:37,960 --> 00:02:41,200
Speaker 3: five dot points worth, and it just feels and smells

47
00:02:41,280 --> 00:02:44,200
Speaker 3: like AI, and it just has all the tales like

48
00:02:44,320 --> 00:02:46,280
Speaker 3: everyone knows about the M dashes. But you can just

49
00:02:46,320 --> 00:02:49,400
Speaker 3: read it and go, ah, this doesn't sound It sounds

50
00:02:49,440 --> 00:02:52,640
Speaker 3: like CHGBT. And so you can get AI to do

51
00:02:52,680 --> 00:02:54,919
Speaker 3: all the work and it sounds like AI. The other

52
00:02:55,000 --> 00:02:57,680
Speaker 3: thing is you can do most of the work and

53
00:02:57,720 --> 00:02:59,680
Speaker 3: then get AI to do a cleanup at the end,

54
00:03:00,360 --> 00:03:02,520
Speaker 3: and it will sound like AI. And so everyone's going

55
00:03:02,560 --> 00:03:04,360
Speaker 3: to look at that and go, the whole thing's written

56
00:03:04,400 --> 00:03:07,640
Speaker 3: by AI, and then they'll disregard it like they probably

57
00:03:07,639 --> 00:03:10,480
Speaker 3: should about the first one, which is just all written

58
00:03:10,600 --> 00:03:14,520
Speaker 3: by AI. So if you build some stuff that sounds

59
00:03:14,560 --> 00:03:17,320
Speaker 3: like AI, people are going to roll their eyes and

60
00:03:17,360 --> 00:03:19,320
Speaker 3: they may not read it with the same kind of

61
00:03:19,360 --> 00:03:21,640
Speaker 3: intent or kind of care that they would if it

62
00:03:21,720 --> 00:03:23,840
Speaker 3: was sounding like you. And they go, oh my god,

63
00:03:23,840 --> 00:03:25,320
Speaker 3: I Mattha right the whole thing, And I really want

64
00:03:25,360 --> 00:03:27,040
Speaker 3: to listen to everything she was said, and I really

65
00:03:27,040 --> 00:03:29,160
Speaker 3: want to read it intensely rather than I don't know

66
00:03:29,160 --> 00:03:32,040
Speaker 3: if AIS written this. Therefore I'll just skim it, or

67
00:03:32,080 --> 00:03:33,440
Speaker 3: I'll get AI to summarize it.

68
00:03:33,919 --> 00:03:36,280
Speaker 1: So the point of this episode is not to say

69
00:03:36,360 --> 00:03:40,520
Speaker 1: don't use AI to help you write things, but perhaps

70
00:03:41,200 --> 00:03:43,320
Speaker 1: don't use AI to do all the thinking and all

71
00:03:43,360 --> 00:03:48,320
Speaker 1: the content creation. And so then the point of us

72
00:03:48,360 --> 00:03:51,640
Speaker 1: helping you not sound like AI means that the work

73
00:03:51,640 --> 00:03:54,720
Speaker 1: that you do produce, that you do genuinely think about

74
00:03:54,800 --> 00:03:59,320
Speaker 1: and work with AI to refine and edit and so forth,

75
00:04:00,120 --> 00:04:02,840
Speaker 1: gets taken a whole lot more seriously rather than people

76
00:04:02,960 --> 00:04:06,240
Speaker 1: just dismissing it. So Neo, I know that in the

77
00:04:06,400 --> 00:04:12,960
Speaker 1: Inventium AI training, when we're teaching people how to use

78
00:04:13,000 --> 00:04:15,760
Speaker 1: AI for writing, you actually do go into this. So

79
00:04:15,800 --> 00:04:18,159
Speaker 1: what are some of the things that you're teaching people

80
00:04:18,520 --> 00:04:22,280
Speaker 1: to avoid sounding like AI?

81
00:04:21,800 --> 00:04:24,600
Speaker 3: The first one I will say is this, Like often

82
00:04:24,600 --> 00:04:26,400
Speaker 3: we'll get it like a paragraph and you might say, look,

83
00:04:26,440 --> 00:04:28,440
Speaker 3: I need you to clean up this paragraph AI, so

84
00:04:28,480 --> 00:04:30,520
Speaker 3: you're not getting it to write the whole damn thing.

85
00:04:31,839 --> 00:04:35,240
Speaker 3: And the first tip is telling AI don't make this

86
00:04:35,279 --> 00:04:39,440
Speaker 3: sound more complicated than the original. So IT often likes

87
00:04:39,480 --> 00:04:41,839
Speaker 3: to have these big words and be very stiff and

88
00:04:41,960 --> 00:04:45,400
Speaker 3: quite formal unless you've previously told it you know what

89
00:04:45,440 --> 00:04:47,120
Speaker 3: you like it to sound like, and things like that.

90
00:04:47,200 --> 00:04:49,960
Speaker 3: So the quickest easiest one is just saying something along

91
00:04:49,960 --> 00:04:52,520
Speaker 3: the lines of, don't make the output more complicated than

92
00:04:52,520 --> 00:04:54,279
Speaker 3: in the regional when you give it a paragraph and

93
00:04:54,320 --> 00:04:56,480
Speaker 3: ask it to critique it and clean it up and

94
00:04:56,520 --> 00:04:59,479
Speaker 3: things like that. But there are other ones that you

95
00:04:59,480 --> 00:05:01,640
Speaker 3: can add in. So the first one is the most

96
00:05:01,640 --> 00:05:03,200
Speaker 3: obvious one, but there's a bit of a catch on

97
00:05:03,279 --> 00:05:05,600
Speaker 3: how to do it. You know those M dashes that

98
00:05:05,640 --> 00:05:08,000
Speaker 3: seem to pop up everywhere, So I'll just explain for

99
00:05:08,040 --> 00:05:10,039
Speaker 3: those who aren't sure what an M dash is, and

100
00:05:10,200 --> 00:05:12,719
Speaker 3: M dash is a longer dash. This is a I

101
00:05:12,720 --> 00:05:14,640
Speaker 3: probably should get you to explain this, you're the actual

102
00:05:14,760 --> 00:05:18,799
Speaker 3: author here, but yeah, it's a proper way to write things, right.

103
00:05:19,160 --> 00:05:22,839
Speaker 1: Yes, yeah, So they're like visually they look like a

104
00:05:22,920 --> 00:05:26,080
Speaker 1: double hyphen So if you can imagine two hyphens connected,

105
00:05:26,160 --> 00:05:29,320
Speaker 1: that's what an m dash looks like. And because AI

106
00:05:29,800 --> 00:05:34,360
Speaker 1: was trained on books, if you read any book and

107
00:05:34,839 --> 00:05:40,279
Speaker 1: they're using you know, like language like and I'm just

108
00:05:40,279 --> 00:05:42,120
Speaker 1: trying to think of a sentence that has like an

109
00:05:42,240 --> 00:05:44,760
Speaker 1: M dash in it, so you know it might be

110
00:05:45,920 --> 00:05:49,240
Speaker 1: run every piece of content through this list before publishing,

111
00:05:51,040 --> 00:05:53,719
Speaker 1: and if you spot three or more, rewrite, so like

112
00:05:53,760 --> 00:05:56,880
Speaker 1: you could theoretically have an M dash before the end,

113
00:05:57,120 --> 00:05:58,600
Speaker 1: just to kind of break up the sentence.

114
00:05:58,640 --> 00:06:00,919
Speaker 2: You'll see it in any book that you're reading.

115
00:06:01,240 --> 00:06:03,560
Speaker 1: And because AI was trained on all those books, a

116
00:06:03,600 --> 00:06:06,280
Speaker 1: lot of M dashes went into the system in terms

117
00:06:06,320 --> 00:06:08,000
Speaker 1: of its knowledge, and it's.

118
00:06:07,800 --> 00:06:11,000
Speaker 3: Really hard baked in there because AI is a product

119
00:06:11,040 --> 00:06:13,839
Speaker 3: of its training, because it was trained on so many books,

120
00:06:14,000 --> 00:06:17,360
Speaker 3: then M dashes are everywhere throughout its training, and it

121
00:06:17,720 --> 00:06:19,800
Speaker 3: just pulls out M dashes an awful lot, and it's

122
00:06:19,839 --> 00:06:22,359
Speaker 3: really hard to get AI to not use M dashes.

123
00:06:22,400 --> 00:06:25,760
Speaker 3: But so rather than just say don't use M dashes,

124
00:06:26,800 --> 00:06:29,240
Speaker 3: I found that that has a very low chance of working.

125
00:06:29,279 --> 00:06:32,599
Speaker 3: Sometimes it does gish kind of Sometimes. The phrase that

126
00:06:32,680 --> 00:06:35,599
Speaker 3: I use, which has given me the best success is this.

127
00:06:36,600 --> 00:06:40,720
Speaker 3: Now it's don't use M dashes, So that is, don't

128
00:06:40,839 --> 00:06:44,640
Speaker 3: use em space dashes. And then I actually give it

129
00:06:44,680 --> 00:06:47,159
Speaker 3: an example of an M dash, So I actually paste

130
00:06:47,160 --> 00:06:49,680
Speaker 3: in an M dash within brackets, so it then knows

131
00:06:49,800 --> 00:06:52,160
Speaker 3: that's an M dash. I've been really clear about this,

132
00:06:53,560 --> 00:06:55,640
Speaker 3: and you could if you don't like short dashes either,

133
00:06:55,720 --> 00:06:57,800
Speaker 3: you could say or short dashes, and then you can

134
00:06:57,839 --> 00:07:01,560
Speaker 3: put the short dash in an inner brackets. Rely on

135
00:07:01,600 --> 00:07:05,200
Speaker 3: commas full stops or short shorter sentences instead, So you

136
00:07:05,200 --> 00:07:06,880
Speaker 3: could do things like that, or you could write it

137
00:07:06,920 --> 00:07:08,960
Speaker 3: if you're affine with the dash, you just want them

138
00:07:09,000 --> 00:07:11,760
Speaker 3: to be short. It'd be don't use M dashes, paste

139
00:07:11,760 --> 00:07:14,160
Speaker 3: in an M dash instead. I want you to use

140
00:07:14,280 --> 00:07:16,720
Speaker 3: short dashes and then paste in the short dash. So

141
00:07:16,720 --> 00:07:18,040
Speaker 3: it's up to you whether you want to ban the

142
00:07:18,120 --> 00:07:21,760
Speaker 3: dash or you just want to keep the short dash,

143
00:07:21,800 --> 00:07:24,559
Speaker 3: but yet give it first off, the words and also

144
00:07:24,640 --> 00:07:26,800
Speaker 3: the example of what the dash is you're talking about,

145
00:07:26,840 --> 00:07:29,800
Speaker 3: and I found that is better, but not one hundred

146
00:07:29,800 --> 00:07:32,520
Speaker 3: percent why it loves the M dash and it's really

147
00:07:32,560 --> 00:07:33,840
Speaker 3: hard to get it to stop doing.

148
00:07:36,240 --> 00:07:38,320
Speaker 1: What are some of the other things that you teach

149
00:07:38,520 --> 00:07:41,840
Speaker 1: in inventium dot AI programs.

150
00:07:42,480 --> 00:07:46,760
Speaker 3: These are all to undo some of the very common

151
00:07:46,840 --> 00:07:49,880
Speaker 3: things that AI does, and with AI doing in the

152
00:07:50,040 --> 00:07:53,160
Speaker 3: commonly things, it's doing it the right way. And I'm

153
00:07:53,200 --> 00:07:56,400
Speaker 3: doing your quotes here because the right way to write

154
00:07:56,560 --> 00:07:58,840
Speaker 3: is the very formal way to write, and it's been

155
00:07:58,840 --> 00:08:00,400
Speaker 3: trained on the right way to write by a whole

156
00:08:00,440 --> 00:08:02,240
Speaker 3: bunch of books and sources, and also you know the

157
00:08:02,240 --> 00:08:05,720
Speaker 3: guides on how to write well, and so therefore you

158
00:08:05,800 --> 00:08:08,480
Speaker 3: know if you're a native English speaker and you're you know,

159
00:08:08,600 --> 00:08:11,800
Speaker 3: you've done a lot of writing, chances are you more

160
00:08:11,840 --> 00:08:15,000
Speaker 3: get caught of these AI detectors and things like that.

161
00:08:15,080 --> 00:08:18,480
Speaker 3: So it's almost like putting the human back into it,

162
00:08:18,600 --> 00:08:21,040
Speaker 3: making it a little bit messier than the perfect way

163
00:08:21,080 --> 00:08:24,280
Speaker 3: to write. Okay, So with that first one's the second

164
00:08:24,280 --> 00:08:26,880
Speaker 3: one is mix sentence lengths. And the reason for this

165
00:08:27,000 --> 00:08:30,600
Speaker 3: is AI generally has the same sentence length again and

166
00:08:30,600 --> 00:08:34,320
Speaker 3: again and again, and so when you and I read this,

167
00:08:34,400 --> 00:08:37,080
Speaker 3: we may not actually pick up that these sentence lengths

168
00:08:37,080 --> 00:08:39,440
Speaker 3: are the same, but we just go this doesn't feel

169
00:08:39,480 --> 00:08:41,200
Speaker 3: like a human so we get that kind of like

170
00:08:41,440 --> 00:08:45,360
Speaker 3: the uncanny valley of writing. So say mix sentence lengths,

171
00:08:45,600 --> 00:08:48,400
Speaker 3: these short ones for punch. He's longer ones when you

172
00:08:48,440 --> 00:08:51,120
Speaker 3: need detail, And I found that one does pretty well.

173
00:08:51,960 --> 00:08:56,000
Speaker 3: So if you remember hending Away famous for short, punchy sentences,

174
00:08:56,160 --> 00:08:58,440
Speaker 3: so get some punchy things in there, get some detailed

175
00:08:58,480 --> 00:09:00,760
Speaker 3: ones that mixes it up and you and I reading

176
00:09:00,760 --> 00:09:04,920
Speaker 3: it go okay. Another one is the language that AI uses.

177
00:09:05,240 --> 00:09:07,599
Speaker 3: It can try to be the smartest AI in the

178
00:09:07,679 --> 00:09:11,120
Speaker 3: room and can like to use those long words as

179
00:09:11,160 --> 00:09:14,520
Speaker 3: part of that don't make the thing more complicated. So

180
00:09:14,559 --> 00:09:17,080
Speaker 3: you could do something like this, use playing language at

181
00:09:17,120 --> 00:09:20,000
Speaker 3: a ninth at about a ninth grade and reading level,

182
00:09:20,880 --> 00:09:25,360
Speaker 3: avoid jargon and buzzwords. Now that the ninth grade thing

183
00:09:25,440 --> 00:09:31,720
Speaker 3: or the fourteen year old type brief is common along

184
00:09:32,080 --> 00:09:34,800
Speaker 3: in industry practice, I understand when the newspaper or pro

185
00:09:34,840 --> 00:09:37,200
Speaker 3: writing and things like that, that's the kind of goal to

186
00:09:37,240 --> 00:09:41,959
Speaker 3: get to. Why because everyone can just read that pretty simply,

187
00:09:42,040 --> 00:09:44,400
Speaker 3: pretty easily. It's not really difficult to read, and you're

188
00:09:44,440 --> 00:09:46,600
Speaker 3: not going to pull out university style words that are

189
00:09:46,640 --> 00:09:49,000
Speaker 3: only a handful of people going to understand. So that's

190
00:09:49,040 --> 00:09:54,040
Speaker 3: why you do that. The other ones are sometimes it

191
00:09:54,120 --> 00:09:57,720
Speaker 3: deals with the way that sentences is constructed. Particularly start

192
00:09:57,720 --> 00:10:00,199
Speaker 3: of sentences can be a bit stiff. So one of

193
00:10:00,240 --> 00:10:02,360
Speaker 3: the things is it's fine to start a sentence with

194
00:10:02,520 --> 00:10:06,360
Speaker 3: and but also although you should do this sparingly, so

195
00:10:06,600 --> 00:10:09,040
Speaker 3: occasionally we will actually start a sentence with a handbuilt

196
00:10:09,080 --> 00:10:10,760
Speaker 3: but or so that's not the right way to do it.

197
00:10:10,760 --> 00:10:14,000
Speaker 3: I'm doing air quotes again, but humans do that, and

198
00:10:14,040 --> 00:10:16,160
Speaker 3: so therefore let's let AI do it. Not do it

199
00:10:16,160 --> 00:10:18,000
Speaker 3: all the time, but we'll let it do it. The

200
00:10:18,080 --> 00:10:22,800
Speaker 3: other one about starting sentences is avoid stiff transitions such

201
00:10:22,840 --> 00:10:28,520
Speaker 3: as additionally and moreover. Who's start human says moreover blah

202
00:10:28,520 --> 00:10:30,920
Speaker 3: blah blah blah blah. You know, you feel like you're

203
00:10:31,040 --> 00:10:33,480
Speaker 3: like a stuffy person in the eighteen hundreds with a

204
00:10:33,520 --> 00:10:36,920
Speaker 3: pipe or something. So I avoid stiff transitions such as

205
00:10:36,920 --> 00:10:41,880
Speaker 3: additionally and moreover. Now this is a weird one, but

206
00:10:41,960 --> 00:10:45,679
Speaker 3: I find this one. I used to like, but now

207
00:10:45,720 --> 00:10:50,560
Speaker 3: I hate that AI does. Often it'll frame things in

208
00:10:50,600 --> 00:10:53,679
Speaker 3: a particular way near the end of a sentence, end

209
00:10:53,679 --> 00:10:57,320
Speaker 3: of a paragraph. It'll say half bunch of explanation stuff,

210
00:10:57,480 --> 00:11:02,520
Speaker 3: and I want to say, it's not this. It's so it

211
00:11:02,520 --> 00:11:06,160
Speaker 3: maybe like, it's not just a good way of writing,

212
00:11:06,880 --> 00:11:10,480
Speaker 3: it's the way to communicate something.

213
00:11:10,240 --> 00:11:12,120
Speaker 2: Like that, right, AI.

214
00:11:13,480 --> 00:11:15,720
Speaker 3: It does this a lot, and actually was at the

215
00:11:15,720 --> 00:11:17,720
Speaker 3: start when it was coming, I was, Oh, that's really clever.

216
00:11:17,800 --> 00:11:20,040
Speaker 3: It does that really well. I like, but it over

217
00:11:20,280 --> 00:11:23,560
Speaker 3: uses that mechanism, and so therefore I actually had to

218
00:11:23,559 --> 00:11:25,760
Speaker 3: google to find out what that mechanism was called. It's

219
00:11:25,760 --> 00:11:31,160
Speaker 3: called corrective antithesis. So I say, never use corrective antithesis,

220
00:11:31,440 --> 00:11:33,880
Speaker 3: And then in brackets, given an example, I say, for example,

221
00:11:34,360 --> 00:11:39,280
Speaker 3: not X, but why and then that certainly cuts that down.

222
00:11:39,880 --> 00:11:42,000
Speaker 3: And there's one more if I can throw it in there,

223
00:11:42,080 --> 00:11:45,400
Speaker 3: which is don't lean on the rule of three. Now,

224
00:11:45,400 --> 00:11:47,640
Speaker 3: the rule of three is generally you give examples. You

225
00:11:47,640 --> 00:11:50,360
Speaker 3: can give three examples, and it kind of makes you know,

226
00:11:50,920 --> 00:11:55,760
Speaker 3: it reads nicely. But AI over uses the rule of three.

227
00:11:56,120 --> 00:11:58,960
Speaker 3: It over uses that in so many things, and so

228
00:11:59,440 --> 00:12:01,559
Speaker 3: I say, don't lean on the rule of three. Use

229
00:12:01,600 --> 00:12:04,199
Speaker 3: one or two items, or use a longer list if

230
00:12:04,200 --> 00:12:07,320
Speaker 3: it strengthens the point. So you can do one or two,

231
00:12:07,520 --> 00:12:09,800
Speaker 3: or you can do five. You know, up to you,

232
00:12:10,120 --> 00:12:14,080
Speaker 3: but don't always use rule of three, because it's that consistency,

233
00:12:14,320 --> 00:12:16,360
Speaker 3: like the sentence lengths being the same, the rule of

234
00:12:16,400 --> 00:12:19,000
Speaker 3: three being the same. It's not this, it's that being

235
00:12:19,080 --> 00:12:22,360
Speaker 3: the same that feels very robotic. Because robots are really

236
00:12:22,400 --> 00:12:24,720
Speaker 3: good at doing things the same ais really good things

237
00:12:24,760 --> 00:12:27,840
Speaker 3: the same. Humans we're messier, and so we're used to

238
00:12:27,880 --> 00:12:29,840
Speaker 3: seeing things that a little bit messier. And it's also

239
00:12:29,920 --> 00:12:31,920
Speaker 3: easier and nicer for us to read because it kind

240
00:12:31,920 --> 00:12:34,240
Speaker 3: of feels more like us. And so this is just

241
00:12:34,280 --> 00:12:36,400
Speaker 3: a whole bunch of points you can put in if

242
00:12:36,480 --> 00:12:39,080
Speaker 3: like you've got this as a standard prompt, all those

243
00:12:39,120 --> 00:12:41,160
Speaker 3: what have I got one to three, four, over, six, seven, eight,

244
00:12:41,520 --> 00:12:44,000
Speaker 3: seven or eight dot points? You can just drop them

245
00:12:44,040 --> 00:12:46,760
Speaker 3: in and then say, when you rewrite my paragraph because

246
00:12:46,800 --> 00:12:49,800
Speaker 3: I needed to be cleaned up, don't do these things.

247
00:12:49,800 --> 00:12:50,880
Speaker 3: That's what you're trying to tell them.

248
00:12:51,520 --> 00:12:52,760
Speaker 2: Okay, I love all those.

249
00:12:52,960 --> 00:12:56,640
Speaker 1: Now what I did ne in preparation for this episode

250
00:12:56,880 --> 00:13:02,880
Speaker 1: is I went into all the llms that I use

251
00:13:04,440 --> 00:13:08,160
Speaker 1: and like I love comparing the output, so I went

252
00:13:08,200 --> 00:13:14,120
Speaker 1: into chatpt Claude, Gemini and also Perplexity, and I gave

253
00:13:14,160 --> 00:13:16,520
Speaker 1: it all the same prompt, which was something along the

254
00:13:16,559 --> 00:13:17,719
Speaker 1: lines of.

255
00:13:19,920 --> 00:13:22,520
Speaker 2: Tell me what are the fifty.

256
00:13:22,520 --> 00:13:27,800
Speaker 1: Most overused or over indexed words and phrases that AI

257
00:13:28,360 --> 00:13:31,360
Speaker 1: likes to use, and I think I suggested to look

258
00:13:31,400 --> 00:13:36,840
Speaker 1: at LinkedIn for some you know, some examples. And then

259
00:13:37,040 --> 00:13:39,840
Speaker 1: what I did with the output that I got from

260
00:13:39,880 --> 00:13:44,200
Speaker 1: each of those four models, I then put them all

261
00:13:44,280 --> 00:13:46,320
Speaker 1: into the one model Claudes generally the one that I

262
00:13:46,400 --> 00:13:49,800
Speaker 1: default too, and I said, give me a summary of

263
00:13:49,880 --> 00:13:53,800
Speaker 1: the ones that are appearing most frequently across these four outputs.

264
00:13:54,120 --> 00:13:58,120
Speaker 1: And so with that, I've then created a bit of

265
00:13:58,160 --> 00:14:01,360
Speaker 1: a mega prompt that people can use that I've called

266
00:14:01,360 --> 00:14:05,480
Speaker 1: the anti ai slop prompt and if you want that,

267
00:14:05,600 --> 00:14:07,240
Speaker 1: I will link to it in the show notes. But

268
00:14:07,280 --> 00:14:09,440
Speaker 1: I'm just going to go through a few things just

269
00:14:09,480 --> 00:14:12,880
Speaker 1: for some additional advice for people that are trying to

270
00:14:13,679 --> 00:14:19,000
Speaker 1: un ai aify their writing so often. And these are

271
00:14:19,000 --> 00:14:22,280
Speaker 1: grouped into some categories. So there's the first category is

272
00:14:22,640 --> 00:14:27,040
Speaker 1: empty opener phrases. So you will have seen things like

273
00:14:27,120 --> 00:14:30,280
Speaker 1: this in today's fast paced world, in an ever changing

274
00:14:30,360 --> 00:14:32,400
Speaker 1: landscape when it comes to.

275
00:14:32,520 --> 00:14:34,200
Speaker 2: Blah, picture this black like.

276
00:14:34,240 --> 00:14:36,440
Speaker 1: These are all sounding really familiar because these are how

277
00:14:36,800 --> 00:14:41,040
Speaker 1: a lot of certainly a lot of LinkedIn post starts start. Now,

278
00:14:41,240 --> 00:14:45,320
Speaker 1: then we've got a category of fake profound hooks, which

279
00:14:45,360 --> 00:14:47,720
Speaker 1: I think, you know, it's not X, it's why is

280
00:14:47,760 --> 00:14:52,200
Speaker 1: certainly one of those. Here's the truth, which is one

281
00:14:52,240 --> 00:14:55,240
Speaker 1: of my pet hates, or sometimes he's the uncomfortable truth,

282
00:14:55,280 --> 00:14:58,880
Speaker 1: he's the inconvenient truth. I hate that one because I

283
00:14:59,040 --> 00:15:02,760
Speaker 1: use that language here the truth or here's the thing.

284
00:15:02,960 --> 00:15:04,800
Speaker 1: I actually use that a lot in my writing. And

285
00:15:04,840 --> 00:15:07,320
Speaker 1: what is really frustrating, as someone who was writing books

286
00:15:07,360 --> 00:15:10,040
Speaker 1: before the age of AI is that I feel like

287
00:15:10,120 --> 00:15:14,080
Speaker 1: I now have to change my writing style, which was

288
00:15:14,640 --> 00:15:17,480
Speaker 1: you know, certainly in some ways similar to what AI

289
00:15:17,640 --> 00:15:20,640
Speaker 1: produces to you know, just added like I used m

290
00:15:20,720 --> 00:15:23,960
Speaker 1: dashes very often, but now I will lean more towards

291
00:15:23,960 --> 00:15:28,960
Speaker 1: brackets and semi collins for like little asides or quips

292
00:15:29,600 --> 00:15:32,680
Speaker 1: as my editor calls them, which is very frustrating when

293
00:15:32,720 --> 00:15:34,160
Speaker 1: you have to change your own writing style so it

294
00:15:34,160 --> 00:15:38,400
Speaker 1: doesn't look like AI. And we've got lots of corporate buzzwords.

295
00:15:38,440 --> 00:15:42,800
Speaker 1: Probably the most common is delve. I was hearing some

296
00:15:42,880 --> 00:15:46,560
Speaker 1: research actually that was done in twenty twenty three where

297
00:15:47,760 --> 00:15:52,440
Speaker 1: academic authors that were most prolific in twenty twenty three

298
00:15:52,440 --> 00:15:55,920
Speaker 1: in terms of getting things published over indexed heavily on

299
00:15:55,960 --> 00:15:58,920
Speaker 1: the word delve. I don't know if you've heard that research.

300
00:15:59,000 --> 00:16:02,360
Speaker 1: I heard Ethan like sharing that recently, which made me laugh.

301
00:16:03,080 --> 00:16:05,560
Speaker 3: I didn't hear that researcher. I'm a bit sad about

302
00:16:05,560 --> 00:16:09,320
Speaker 3: delve because I use delve as well, but AI uses

303
00:16:09,360 --> 00:16:12,840
Speaker 3: it more. And so you may add a bunch of words.

304
00:16:12,880 --> 00:16:14,720
Speaker 3: Can we also talk about other words we may want

305
00:16:14,760 --> 00:16:16,320
Speaker 3: to add to that banned list?

306
00:16:17,360 --> 00:16:19,760
Speaker 2: For sure? For sure? Give me like three.

307
00:16:19,720 --> 00:16:24,680
Speaker 3: More quietly is another one that I was going.

308
00:16:24,480 --> 00:16:25,720
Speaker 2: To say quietly.

309
00:16:26,040 --> 00:16:29,280
Speaker 1: So the next category actually is what AI has termed

310
00:16:29,400 --> 00:16:34,040
Speaker 1: hollow adjectives and nouns and adjectives, and if you don't

311
00:16:34,080 --> 00:16:38,720
Speaker 1: remember from your English class, and adjective is a descriptive

312
00:16:38,760 --> 00:16:42,000
Speaker 1: word for a noun, So it describes a noun which

313
00:16:42,040 --> 00:16:46,000
Speaker 1: is like an object or a person, a person, place

314
00:16:46,120 --> 00:16:48,760
Speaker 1: or thing. I think is how my English teacher probably

315
00:16:48,760 --> 00:16:54,720
Speaker 1: would have described it. Quietly is so overused. And again

316
00:16:54,760 --> 00:16:56,800
Speaker 1: that really bothers me because that is a word that

317
00:16:56,840 --> 00:16:58,440
Speaker 1: I would have used, but now I feel like I

318
00:16:58,480 --> 00:17:00,280
Speaker 1: have to strip it out mostly from my writing.

319
00:17:00,360 --> 00:17:02,880
Speaker 2: Okay, what are the words near this?

320
00:17:03,680 --> 00:17:06,600
Speaker 3: Ultimately, I would recommend you build your own list of

321
00:17:06,640 --> 00:17:08,920
Speaker 3: words that you don't want AI to use, because there's

322
00:17:08,920 --> 00:17:11,199
Speaker 3: some there that AI will have used, like Delvin quietly,

323
00:17:11,560 --> 00:17:13,760
Speaker 3: but there may be others that IT would use that

324
00:17:14,160 --> 00:17:17,679
Speaker 3: I find words but you would never use. So I

325
00:17:17,800 --> 00:17:21,520
Speaker 3: use the example of like buzzword bingo, so like synergistic. Right,

326
00:17:21,720 --> 00:17:24,360
Speaker 3: if AI uses the words synergistic and you would never

327
00:17:24,480 --> 00:17:27,359
Speaker 3: use it, then put a list together and add to

328
00:17:27,400 --> 00:17:29,719
Speaker 3: that list, but sudogistic in there, maybe put delve in there,

329
00:17:29,720 --> 00:17:32,080
Speaker 3: maybe put quietly in there, and go, don't ever use

330
00:17:32,119 --> 00:17:34,840
Speaker 3: these words, And so build up that list and enhance

331
00:17:34,920 --> 00:17:37,640
Speaker 3: that over time as AI uses words that just great

332
00:17:37,720 --> 00:17:39,920
Speaker 3: on you, because if you've been grading on you, they

333
00:17:39,920 --> 00:17:40,760
Speaker 3: don't sound like you.

334
00:17:41,480 --> 00:17:42,240
Speaker 2: So hopefully.

335
00:17:42,280 --> 00:17:44,360
Speaker 1: I mean, this is so much to take in, as

336
00:17:44,400 --> 00:17:47,640
Speaker 1: I said, for those that are listening and like, yeah,

337
00:17:47,720 --> 00:17:50,880
Speaker 1: I really, I really need to improve my writing output

338
00:17:50,920 --> 00:17:53,159
Speaker 1: because I could relate to a lot of those things,

339
00:17:53,240 --> 00:17:55,399
Speaker 1: and I think about my writing. Yes, I'm using a

340
00:17:55,400 --> 00:17:56,760
Speaker 1: lot of the things that Neo and a manter I

341
00:17:56,760 --> 00:17:59,919
Speaker 1: have talked about. Go to the show notes. There is

342
00:18:00,240 --> 00:18:04,720
Speaker 1: a link to get our anti AI slopped prompt, which

343
00:18:04,800 --> 00:18:07,200
Speaker 1: I highly recommend you customize.

344
00:18:07,240 --> 00:18:08,960
Speaker 2: So it's a lengthy prompt.

345
00:18:09,880 --> 00:18:12,000
Speaker 1: It includes a lot of what we've spoken about today,

346
00:18:12,040 --> 00:18:16,000
Speaker 1: but do read it and put in your own personal

347
00:18:16,880 --> 00:18:19,480
Speaker 1: pet peeves so that you can customize it and run

348
00:18:19,560 --> 00:18:23,119
Speaker 1: any content that you create or create with AI through

349
00:18:23,359 --> 00:18:26,479
Speaker 1: this prompt. It will improve your work a lot. And

350
00:18:26,680 --> 00:18:30,200
Speaker 1: also I would say, please share this episode with maybe

351
00:18:30,240 --> 00:18:33,639
Speaker 1: some people that you know that are perhaps producing a

352
00:18:33,640 --> 00:18:35,760
Speaker 1: bit of AI slot, but they are actually doing the

353
00:18:35,800 --> 00:18:38,879
Speaker 1: thinking themselves, which is the worst combination where the human

354
00:18:38,960 --> 00:18:41,600
Speaker 1: is doing the majority of the work and AI is

355
00:18:41,640 --> 00:18:44,520
Speaker 1: then going over it and AI are fying it and

356
00:18:44,560 --> 00:18:48,280
Speaker 1: completely destroying that person's credibility, so do share this episode

357
00:18:48,359 --> 00:18:53,119
Speaker 1: with those that need it. How i AI was hosted

358
00:18:53,119 --> 00:18:56,200
Speaker 1: by me Amantha Imber and Neo Applan. A big thank

359
00:18:56,200 --> 00:18:58,760
Speaker 1: you to Martin Imba who does our sound editing and

360
00:18:58,880 --> 00:19:02,280
Speaker 1: Jem Rubio for production support, and thank you to John

361
00:19:02,359 --> 00:19:04,320
Speaker 1: Kilby who composed the theme music