1
00:00:00,040 --> 00:00:02,840
Speaker 1: Hey everyone, it's Robert and Joe here. Today we've got

2
00:00:02,840 --> 00:00:05,080
Speaker 1: something a little different to share with you. It's a

3
00:00:05,120 --> 00:00:08,719
Speaker 1: new season of the Smart Talks with IBM podcast series.

4
00:00:09,200 --> 00:00:12,000
Speaker 2: This season, on smart Talks, Malcolm Gladwell and team are

5
00:00:12,039 --> 00:00:15,240
Speaker 2: diving into the transformative world of artificial intelligence with a

6
00:00:15,280 --> 00:00:18,600
Speaker 2: fresh perspective on the concept of open What does open

7
00:00:18,680 --> 00:00:21,880
Speaker 2: really mean in the context of AI. It can mean

8
00:00:22,000 --> 00:00:25,599
Speaker 2: open source code or open data, but it also encompasses

9
00:00:25,680 --> 00:00:30,760
Speaker 2: fostering an ecosystem of ideas, ensuring diverse perspectives are heard,

10
00:00:31,120 --> 00:00:33,519
Speaker 2: and enabling new levels of transparency.

11
00:00:33,840 --> 00:00:37,080
Speaker 1: Join hosts from your favorite pushkin podcasts as they explore

12
00:00:37,080 --> 00:00:40,919
Speaker 1: how openness in AI is reshaping industries, driving innovation, and

13
00:00:40,960 --> 00:00:44,840
Speaker 1: redefining what's possible. You'll hear from industry experts and leaders

14
00:00:44,840 --> 00:00:48,320
Speaker 1: about the implication and possibilities of open AI, and of course,

15
00:00:48,680 --> 00:00:50,880
Speaker 1: Malcolm Gladwell will be there to guide you through the

16
00:00:50,920 --> 00:00:52,680
Speaker 1: season with his unique insights.

17
00:00:52,960 --> 00:00:55,520
Speaker 2: Look out for new episodes of Smart Talks every other

18
00:00:55,600 --> 00:00:59,160
Speaker 2: week on the iHeartRadio app, Apple Podcasts, or wherever you

19
00:00:59,200 --> 00:01:02,480
Speaker 2: get your podcast and learn more at IBM dot com

20
00:01:02,520 --> 00:01:03,840
Speaker 2: slash smart Talks.

21
00:01:04,480 --> 00:01:07,479
Speaker 3: Hey Malcolm Glawell here I'm back in your feed today

22
00:01:07,560 --> 00:01:10,399
Speaker 3: because we are re releasing an episode of Smart Talks

23
00:01:10,400 --> 00:01:15,480
Speaker 3: with IBM on a very timely topic, AI governance and

24
00:01:15,560 --> 00:01:20,399
Speaker 3: why regulation is critical to building responsible and accountable AI.

25
00:01:20,920 --> 00:01:26,759
Speaker 3: I hope you enjoy it. Hello, Hello, Welcome to Smart

26
00:01:26,760 --> 00:01:32,440
Speaker 3: Talks with IBM, a podcast from Pushkin Industries, iHeartRadio and IBM.

27
00:01:32,760 --> 00:01:37,120
Speaker 3: I'm Malcolm Glabwell. This season, we're continuing our conversation with

28
00:01:37,319 --> 00:01:41,800
Speaker 3: new creators visionaries who are creatively applying technology in business

29
00:01:41,800 --> 00:01:45,520
Speaker 3: to drive change, but with a focus on the transformative

30
00:01:45,560 --> 00:01:49,480
Speaker 3: power of artificial intelligence and what it means to leverage

31
00:01:49,520 --> 00:01:54,360
Speaker 3: AI as a game changing multiplier for your business. Our

32
00:01:54,400 --> 00:01:59,600
Speaker 3: guest today is Christina Montgomery, IBM's Chief Privacy and Trust Officer.

33
00:02:00,200 --> 00:02:04,160
Speaker 3: She's also chair of IBM's AI Ethics Board. In addition

34
00:02:04,280 --> 00:02:08,639
Speaker 3: to overseeing IBM's privacy policy, A core part of Christina's

35
00:02:08,680 --> 00:02:13,080
Speaker 3: job involves AI governance, making sure the way AI is

36
00:02:13,200 --> 00:02:19,440
Speaker 3: used complies with the international legal regulations customized for each industry.

37
00:02:20,160 --> 00:02:25,320
Speaker 3: In today's episode, Christina will explain why businesses need foundational

38
00:02:25,360 --> 00:02:29,920
Speaker 3: principles when it comes to using technology, why AI regulation

39
00:02:30,280 --> 00:02:34,600
Speaker 3: should focus on specific use cases over the technology itself

40
00:02:35,120 --> 00:02:39,280
Speaker 3: and share a little bit about her landmark congressional testimony.

41
00:02:39,680 --> 00:02:43,880
Speaker 3: Last May, Christina spoke with doctor Lori Santos, host of

42
00:02:43,919 --> 00:02:48,640
Speaker 3: the Pushkin podcast The Happiness Lab, a cognitive scientist and

43
00:02:48,880 --> 00:02:53,160
Speaker 3: psychology professor at Yale University. Laurie is an expert on

44
00:02:53,400 --> 00:03:00,160
Speaker 3: human happiness and cognition. Okay, let's get to the interview.

45
00:03:01,040 --> 00:03:03,120
Speaker 4: So, Christina, I'm so excited to talk to you today.

46
00:03:03,280 --> 00:03:05,480
Speaker 4: So let's start by talking a little bit about your

47
00:03:05,560 --> 00:03:08,239
Speaker 4: role at IBM. What does a Chief Privacy and Trust

48
00:03:08,280 --> 00:03:09,320
Speaker 4: Officer actually do.

49
00:03:10,120 --> 00:03:13,880
Speaker 5: It's a really dynamic profession and it's not a new profession,

50
00:03:14,440 --> 00:03:17,280
Speaker 5: but the role has really changed. I mean, my role

51
00:03:17,440 --> 00:03:21,120
Speaker 5: today is broader than just helping to ensure compliance with

52
00:03:21,280 --> 00:03:25,920
Speaker 5: data protection laws globally. I'm also responsible for AI governance

53
00:03:25,960 --> 00:03:28,880
Speaker 5: I co chair or AI Ethics Board here at IBM,

54
00:03:29,240 --> 00:03:32,280
Speaker 5: and for data clearance and data governance as well for

55
00:03:32,360 --> 00:03:35,920
Speaker 5: the company. So I have both a compliance aspect to

56
00:03:35,960 --> 00:03:38,680
Speaker 5: my role, really important on a global basis, but also

57
00:03:39,200 --> 00:03:44,120
Speaker 5: help the business to competitively differentiate because really trust is

58
00:03:44,120 --> 00:03:47,840
Speaker 5: a strategic advantage for IBM and a competitive differentiator as

59
00:03:47,840 --> 00:03:51,760
Speaker 5: a company that's been responsibly managing the most sensitive data

60
00:03:51,800 --> 00:03:54,400
Speaker 5: for our clients for more than a century now and

61
00:03:54,560 --> 00:03:57,200
Speaker 5: helping to usher new technologies into the world with trust

62
00:03:57,280 --> 00:04:00,680
Speaker 5: and transparency, and so that's also a key aspect of

63
00:04:00,720 --> 00:04:01,119
Speaker 5: my role.

64
00:04:01,640 --> 00:04:03,840
Speaker 4: And so you joined us here on smart Talks back

65
00:04:03,840 --> 00:04:06,080
Speaker 4: in twenty twenty one and you chatted with us about

66
00:04:06,080 --> 00:04:09,760
Speaker 4: IBM's approach of building trust and transparency with AI, and

67
00:04:09,920 --> 00:04:11,960
Speaker 4: that was only two years ago. But it almost feels

68
00:04:11,960 --> 00:04:14,280
Speaker 4: like in eternity has happened in the field of AI

69
00:04:14,440 --> 00:04:17,440
Speaker 4: since then, and so I'm curious how much has changed

70
00:04:17,480 --> 00:04:19,960
Speaker 4: since you were here last time. Were the things you

71
00:04:20,000 --> 00:04:22,680
Speaker 4: told us before, you know, are they still true? How

72
00:04:22,720 --> 00:04:23,520
Speaker 4: are things changing?

73
00:04:24,200 --> 00:04:27,640
Speaker 5: You're absolutely right, it feels like the world has changed

74
00:04:27,760 --> 00:04:31,440
Speaker 5: really in the last two years. But the same fundamental

75
00:04:31,480 --> 00:04:36,400
Speaker 5: principles and the same overall governance applied to IBM's program

76
00:04:36,960 --> 00:04:41,320
Speaker 5: for data protection and responsible AI that we talked about

77
00:04:41,360 --> 00:04:44,039
Speaker 5: two years ago, and not much has changed there from

78
00:04:44,080 --> 00:04:47,240
Speaker 5: our perspective. And the good thing is we've put these

79
00:04:47,279 --> 00:04:51,719
Speaker 5: practices and this governance approach into place, and we've have

80
00:04:51,960 --> 00:04:55,600
Speaker 5: an established way of looking at these emerging technologies. As

81
00:04:55,640 --> 00:04:59,040
Speaker 5: the technology evolves, the tech is more powerful, for sure.

82
00:04:59,200 --> 00:05:03,080
Speaker 5: Foundation model as are vastly larger and more capable and

83
00:05:03,120 --> 00:05:06,479
Speaker 5: are creating in some respects new issues. But that just

84
00:05:06,520 --> 00:05:08,360
Speaker 5: makes it all the more urgent to do what we've

85
00:05:08,400 --> 00:05:11,279
Speaker 5: been doing and to put trust and transparency into place

86
00:05:11,640 --> 00:05:14,840
Speaker 5: across the business to be accountable to those principles.

87
00:05:15,800 --> 00:05:17,960
Speaker 4: And so our conversation today is really centered around this

88
00:05:18,120 --> 00:05:21,359
Speaker 4: need for new AI regulation and part of that regulation

89
00:05:21,480 --> 00:05:24,359
Speaker 4: involves the mitigation of bias. And this is something I

90
00:05:24,360 --> 00:05:26,719
Speaker 4: think about a ton as a psychologist, right, you know,

91
00:05:26,760 --> 00:05:29,520
Speaker 4: I know, like my students and everyone who's interacting with

92
00:05:29,600 --> 00:05:32,680
Speaker 4: AI is assuming that the kind of knowledge that they're

93
00:05:32,720 --> 00:05:35,719
Speaker 4: getting from this kind of learning is accurate, right, But

94
00:05:35,800 --> 00:05:37,840
Speaker 4: of course AI is only as good as the knowledge

95
00:05:37,839 --> 00:05:40,120
Speaker 4: that's going in. And so talk to me a little

96
00:05:40,160 --> 00:05:43,400
Speaker 4: bit about like why bias occurs in AI and the

97
00:05:43,480 --> 00:05:45,360
Speaker 4: level of the problem that we're really dealing with.

98
00:05:46,480 --> 00:05:49,800
Speaker 5: Yeah, Well, obviously AI is based on data, rights it's

99
00:05:49,880 --> 00:05:54,880
Speaker 5: trained with data, and that data could be biased in

100
00:05:54,880 --> 00:05:57,360
Speaker 5: and of itself, and that's where issues could come up.

101
00:05:57,400 --> 00:05:59,280
Speaker 5: They come up in the data, they could also come

102
00:05:59,360 --> 00:06:02,880
Speaker 5: up in the output of the models themselves. So it's

103
00:06:03,000 --> 00:06:07,760
Speaker 5: really important that you build bias consideration and bias testing

104
00:06:07,880 --> 00:06:11,280
Speaker 5: into your product development cycle, and so what we've been

105
00:06:11,320 --> 00:06:13,880
Speaker 5: thinking about here at IBM and doing We had some

106
00:06:13,920 --> 00:06:16,880
Speaker 5: of our research teams delivered some of the very first

107
00:06:16,920 --> 00:06:20,279
Speaker 5: toolkits to help detect bias years ago now right and

108
00:06:20,279 --> 00:06:23,960
Speaker 5: deployed them to open source. And we have put into

109
00:06:24,000 --> 00:06:27,320
Speaker 5: place for our developers here at IBM and Ethics by

110
00:06:27,320 --> 00:06:30,560
Speaker 5: Design playbook that's sort of a step by step approach

111
00:06:31,320 --> 00:06:37,839
Speaker 5: which also addresses very fully bias considerations, and we provide

112
00:06:37,880 --> 00:06:41,080
Speaker 5: not only like here's a point when you should test

113
00:06:41,120 --> 00:06:43,480
Speaker 5: for it and you consider it in the data, you

114
00:06:43,560 --> 00:06:45,280
Speaker 5: have to measure it both at the data and the

115
00:06:45,320 --> 00:06:48,840
Speaker 5: model level or the outcome level, and we provide guidance

116
00:06:48,839 --> 00:06:51,520
Speaker 5: with respect to what tools can best be used to

117
00:06:51,560 --> 00:06:55,039
Speaker 5: accomplish that. So it's a really important issue. It's one

118
00:06:55,080 --> 00:06:58,080
Speaker 5: you can't just talk about. You have to provide essentially

119
00:06:58,120 --> 00:07:01,760
Speaker 5: the technology and the capabilities of guidance to enable people

120
00:07:01,800 --> 00:07:02,600
Speaker 5: to test for it.

121
00:07:03,040 --> 00:07:05,720
Speaker 4: Recently, you had this wonderful opportunity to head to Congress

122
00:07:05,760 --> 00:07:08,799
Speaker 4: to talk about AI, and in your testimony before Congress

123
00:07:08,839 --> 00:07:11,840
Speaker 4: you mentioned that it's often said that innovation moves too

124
00:07:11,880 --> 00:07:14,480
Speaker 4: fast for government to keep up, and this is something

125
00:07:14,480 --> 00:07:16,640
Speaker 4: that I also worry about as a psychologist, right our

126
00:07:16,680 --> 00:07:19,600
Speaker 4: policy makers really understanding the issues that they're dealing with,

127
00:07:20,040 --> 00:07:22,240
Speaker 4: and so I'm curious how you're approaching this challenge of

128
00:07:22,320 --> 00:07:24,760
Speaker 4: adapting AI policies to keep up with the sort of

129
00:07:24,880 --> 00:07:27,600
Speaker 4: rapid pace of all the advancements we're seeing in AI

130
00:07:27,680 --> 00:07:28,760
Speaker 4: technology itself.

131
00:07:29,960 --> 00:07:33,160
Speaker 5: I think it's really critically important that you have foundational

132
00:07:33,240 --> 00:07:38,680
Speaker 5: principles that applied to not only how you use technology,

133
00:07:38,720 --> 00:07:40,200
Speaker 5: but whether you're going to use it in the first

134
00:07:40,240 --> 00:07:41,840
Speaker 5: place and where you're going to use and apply it

135
00:07:41,880 --> 00:07:45,960
Speaker 5: across your company. And then your program from a governance perspective,

136
00:07:46,040 --> 00:07:48,560
Speaker 5: has to be agile. It has to be able to

137
00:07:48,640 --> 00:07:54,320
Speaker 5: address emerging capabilities, new training methods, etc. And part of

138
00:07:54,360 --> 00:07:58,920
Speaker 5: that involves helping to educate and instill and empower a

139
00:07:58,960 --> 00:08:02,760
Speaker 5: trustworthy culture at a company so you can spot those

140
00:08:02,760 --> 00:08:05,040
Speaker 5: issues so you can ask the right questions at the

141
00:08:05,120 --> 00:08:08,320
Speaker 5: right time if you try. We talked about during the

142
00:08:08,360 --> 00:08:12,560
Speaker 5: Senate hearing, and IBM's been talking for years about regulating

143
00:08:12,880 --> 00:08:16,040
Speaker 5: the use, not the technology itself, because if you try

144
00:08:16,040 --> 00:08:19,880
Speaker 5: to regulate technology, you're very quickly going to find out

145
00:08:20,280 --> 00:08:23,160
Speaker 5: regulation will absolutely never keep up with that.

146
00:08:23,440 --> 00:08:25,680
Speaker 4: And so in your testimony to Congress. You also talked

147
00:08:25,680 --> 00:08:28,920
Speaker 4: about this idea of a precision regulation approach for AI.

148
00:08:29,520 --> 00:08:31,720
Speaker 4: Tell me more about this. What is a precision regulation

149
00:08:31,800 --> 00:08:34,160
Speaker 4: approach and why could that be so important.

150
00:08:34,360 --> 00:08:37,280
Speaker 5: It's funny because I was able to share with Congress

151
00:08:37,960 --> 00:08:41,760
Speaker 5: our precision regulation point of view in twenty twenty three,

152
00:08:42,080 --> 00:08:44,960
Speaker 5: but that precision regulation point of view was published by

153
00:08:45,040 --> 00:08:49,720
Speaker 5: IBM in twenty twenty. So we have not changed our

154
00:08:49,840 --> 00:08:54,720
Speaker 5: position that you should apply the tightest controls, the strictest

155
00:08:54,760 --> 00:08:59,040
Speaker 5: regulatory requirements to the technology where the end use and

156
00:08:59,160 --> 00:09:03,080
Speaker 5: risk of societal harm is the greatest. So that's essentially

157
00:09:03,120 --> 00:09:06,080
Speaker 5: what it is. There's lots of AI technology that's used

158
00:09:06,120 --> 00:09:09,760
Speaker 5: today that doesn't touch people, that's very low risk in nature.

159
00:09:10,280 --> 00:09:13,680
Speaker 5: And even when you think about AI that delivers a

160
00:09:13,840 --> 00:09:19,280
Speaker 5: movie recommendation versus AI that is used to diagnose cancer, right,

161
00:09:19,320 --> 00:09:22,800
Speaker 5: there's very different implications associated with those two uses of

162
00:09:22,840 --> 00:09:27,640
Speaker 5: the technology. And so essentially what precision regulation is is

163
00:09:27,679 --> 00:09:31,079
Speaker 5: apply different rules to different risks, right, more stringent regulation

164
00:09:31,360 --> 00:09:34,480
Speaker 5: to the use cases with the greatest risk. And then

165
00:09:34,640 --> 00:09:38,800
Speaker 5: also we build that out calling for things like transparency

166
00:09:39,400 --> 00:09:42,960
Speaker 5: you see it today with content right, misinformation and the like.

167
00:09:43,559 --> 00:09:47,479
Speaker 5: We believe that consumers should always know when they're interacting

168
00:09:47,520 --> 00:09:50,440
Speaker 5: with an AI system, so be transparent, don't hydrar AI

169
00:09:51,559 --> 00:09:55,160
Speaker 5: clearly define the risks. So as a country, we need

170
00:09:55,200 --> 00:09:58,240
Speaker 5: to have some clear guidance right in globally as well

171
00:09:58,600 --> 00:10:02,200
Speaker 5: in terms of which uses of AI are higher risk

172
00:10:02,280 --> 00:10:06,600
Speaker 5: where we'll apply higher and stricter regulation and have sort

173
00:10:06,640 --> 00:10:09,560
Speaker 5: of a common understanding of what those high risk uses

174
00:10:09,600 --> 00:10:13,160
Speaker 5: are and then demonstrate the impact in the cases of

175
00:10:13,200 --> 00:10:17,960
Speaker 5: those higher risk uses. So companies who are using AI

176
00:10:18,160 --> 00:10:22,520
Speaker 5: in spaces where they can impact people's legal rights, for example,

177
00:10:23,120 --> 00:10:27,319
Speaker 5: should have to conduct an impact assessment that demonstrates that

178
00:10:27,400 --> 00:10:30,600
Speaker 5: the technology isn't biased. So we've been pretty clear about

179
00:10:31,040 --> 00:10:34,760
Speaker 5: apply the most stringent regulation to the highest risk uses

180
00:10:34,760 --> 00:10:35,160
Speaker 5: of AI.

181
00:10:36,600 --> 00:10:38,880
Speaker 4: And so so far we've been talking about your congressional

182
00:10:38,880 --> 00:10:41,720
Speaker 4: testimony in terms of, you know, the specific content that

183
00:10:41,800 --> 00:10:44,240
Speaker 4: you talked about, But I'm just curious on a personal level,

184
00:10:44,559 --> 00:10:46,679
Speaker 4: you know, what was that like right like, right now

185
00:10:46,679 --> 00:10:48,959
Speaker 4: it feels like at a policy level, like there's a

186
00:10:49,040 --> 00:10:51,400
Speaker 4: kind of fever pitch going on with AI right now.

187
00:10:51,640 --> 00:10:53,080
Speaker 4: You know what did that feel like to kind of

188
00:10:53,160 --> 00:10:55,600
Speaker 4: really have the opportunity to talk to policy makers and

189
00:10:55,640 --> 00:10:58,640
Speaker 4: sort of influence what they're thinking about AI technologies like

190
00:10:58,720 --> 00:10:59,640
Speaker 4: in the coming century.

191
00:10:59,679 --> 00:11:02,920
Speaker 5: Perhaps I was really an honor to be able to

192
00:11:03,000 --> 00:11:05,960
Speaker 5: do that and to be one of the first set

193
00:11:06,000 --> 00:11:10,080
Speaker 5: of invitees to the first hearing. And what I learned

194
00:11:10,120 --> 00:11:13,280
Speaker 5: from it essentially is, you know, really two things. The

195
00:11:13,320 --> 00:11:17,199
Speaker 5: first is really the value of authenticity. So both as

196
00:11:17,360 --> 00:11:21,320
Speaker 5: an individual and as a company, I was able to

197
00:11:21,360 --> 00:11:23,800
Speaker 5: talk about what I do. You know, I didn't need

198
00:11:23,840 --> 00:11:27,120
Speaker 5: a lot of advanced prep right, I talked about what

199
00:11:27,280 --> 00:11:30,720
Speaker 5: my job is, what IBM has been putting in place

200
00:11:30,800 --> 00:11:35,000
Speaker 5: for years now. So this isn't about creating something. This

201
00:11:35,160 --> 00:11:37,760
Speaker 5: was just about showing up and being authentic. And we

202
00:11:37,760 --> 00:11:40,920
Speaker 5: were invited for a reason. We were invited because we

203
00:11:40,920 --> 00:11:44,800
Speaker 5: were one of the earliest companies in the AI technology space.

204
00:11:45,440 --> 00:11:50,559
Speaker 5: We're the oldest technology company and we are trusted, and

205
00:11:50,920 --> 00:11:53,320
Speaker 5: that's an honor. And then the second thing I came

206
00:11:53,320 --> 00:11:56,040
Speaker 5: away with was really how important this issue is to society.

207
00:11:56,120 --> 00:12:00,600
Speaker 5: I don't think I appreciated it as much until following

208
00:12:00,720 --> 00:12:05,040
Speaker 5: that experience. I had outreached from colleagues I hadn't worked

209
00:12:05,080 --> 00:12:07,920
Speaker 5: with for years. I had an outreach from family members

210
00:12:07,960 --> 00:12:10,640
Speaker 5: who heard me on the radio. You know, my mother

211
00:12:10,800 --> 00:12:13,679
Speaker 5: and my mother in law, and my nieces and nephews

212
00:12:13,720 --> 00:12:16,080
Speaker 5: and my friends of my kids were all like, Oh,

213
00:12:16,120 --> 00:12:18,079
Speaker 5: I get it. I get what you do. Now, Wow,

214
00:12:18,120 --> 00:12:21,400
Speaker 5: that's pretty cool, you know. So that was really probably

215
00:12:21,400 --> 00:12:24,160
Speaker 5: the best and most impactful takeaway that I had.

216
00:12:24,400 --> 00:12:28,040
Speaker 3: The mass adoption of generative AI, happening at breakneck speed,

217
00:12:28,400 --> 00:12:32,440
Speaker 3: has spurred societies and governments around the world to get

218
00:12:32,520 --> 00:12:38,800
Speaker 3: serious about regulating AI. For businesses, compliance is complex enough already,

219
00:12:39,000 --> 00:12:42,320
Speaker 3: but throw aever involving technology like AI into the mix,

220
00:12:42,720 --> 00:12:49,000
Speaker 3: and compliance itself becomes an exercise in adaptability. As regulators

221
00:12:49,000 --> 00:12:53,559
Speaker 3: seek greater accountability in how AI is used, businesses need

222
00:12:53,640 --> 00:12:58,560
Speaker 3: help creating governance processes comprehensive enough to comply with the law,

223
00:12:58,800 --> 00:13:01,880
Speaker 3: but agile enough to keep up with the rapid rate

224
00:13:01,920 --> 00:13:07,280
Speaker 3: of change in AI development. Regulatory scrutiny isn't the only

225
00:13:07,360 --> 00:13:12,520
Speaker 3: consideration either responsible AI governance. A business's ability to prove

226
00:13:12,559 --> 00:13:17,400
Speaker 3: its AI models are transparent and explainable is also key

227
00:13:17,440 --> 00:13:22,480
Speaker 3: to building trust with customers, regardless of industry. In the

228
00:13:22,520 --> 00:13:26,720
Speaker 3: next part of their conversation, Laurie asked Christina what businesses

229
00:13:26,760 --> 00:13:31,160
Speaker 3: should consider when approaching AI governance. Let's listen.

230
00:13:32,000 --> 00:13:34,440
Speaker 4: So what's a particular role that businesses are playing in

231
00:13:34,520 --> 00:13:37,319
Speaker 4: AI governance? Like why is it so critical for businesses

232
00:13:37,360 --> 00:13:38,280
Speaker 4: to be part of this?

233
00:13:39,160 --> 00:13:43,680
Speaker 5: So I think it's really critically important that businesses understand

234
00:13:44,160 --> 00:13:47,079
Speaker 5: the impacts that technology can have, both in making them

235
00:13:47,080 --> 00:13:50,920
Speaker 5: better businesses, but the impacts that those technologies can have

236
00:13:51,400 --> 00:13:56,520
Speaker 5: on the consumers that they are supporting. You know, businesses

237
00:13:56,640 --> 00:14:01,280
Speaker 5: need to be deploying AI technology that is in alignment

238
00:14:01,400 --> 00:14:03,280
Speaker 5: with the goals that they set for it, and that

239
00:14:03,400 --> 00:14:06,280
Speaker 5: can be trusted. I think for us and for our clients,

240
00:14:06,600 --> 00:14:09,520
Speaker 5: a lot of this comes back to trust in tech.

241
00:14:09,920 --> 00:14:15,640
Speaker 5: If you deploy something that doesn't work, that hallucinates, that discriminates,

242
00:14:16,120 --> 00:14:20,280
Speaker 5: that isn't transparent, where decisions can't be explained, then you

243
00:14:20,400 --> 00:14:24,040
Speaker 5: are going to very rapidly erode the trust at best

244
00:14:24,160 --> 00:14:27,520
Speaker 5: right of your clients and at worst for yourself. You're

245
00:14:27,520 --> 00:14:30,200
Speaker 5: going to create legal and regulatory issues for yourself as well.

246
00:14:30,280 --> 00:14:34,200
Speaker 5: So trusted technology is really important, and I think there's

247
00:14:34,240 --> 00:14:36,440
Speaker 5: a lot of pressure on businesses today to move very

248
00:14:36,520 --> 00:14:39,200
Speaker 5: rapidly and adopt technology. But if you do it without

249
00:14:39,280 --> 00:14:42,760
Speaker 5: having a program of governance in place, you're really risking

250
00:14:42,960 --> 00:14:44,440
Speaker 5: eroding that trust.

251
00:14:44,400 --> 00:14:46,440
Speaker 4: And so this is really where I think a strong

252
00:14:46,480 --> 00:14:50,040
Speaker 4: AI governance comes in. Talk about from your perspective, how

253
00:14:50,280 --> 00:14:53,640
Speaker 4: this really contributes to maintaining the trust that customers and

254
00:14:53,680 --> 00:14:55,560
Speaker 4: stakeholders have in these technologies.

255
00:14:55,760 --> 00:14:58,320
Speaker 5: Yeah. Absolutely, I mean you need to have a governance

256
00:14:58,360 --> 00:15:01,840
Speaker 5: program because you need to understand and that the technology,

257
00:15:01,920 --> 00:15:06,920
Speaker 5: particularly in the AI space, that you are deploying, is explainable.

258
00:15:07,000 --> 00:15:11,320
Speaker 5: You need to understand why it's making decisions and recommendations

259
00:15:11,400 --> 00:15:12,880
Speaker 5: that it's making, and you need to be able to

260
00:15:12,880 --> 00:15:15,280
Speaker 5: explain that to your consumers. I mean, you can't do

261
00:15:15,360 --> 00:15:17,440
Speaker 5: that if you don't know where your data is coming from,

262
00:15:17,440 --> 00:15:19,920
Speaker 5: what data are you using to train those models, if

263
00:15:19,920 --> 00:15:23,720
Speaker 5: you don't have a program that manages the alignment of

264
00:15:23,760 --> 00:15:27,320
Speaker 5: your AI models over time to make sure as AI

265
00:15:27,520 --> 00:15:31,960
Speaker 5: learns and evolves over uses, which is in large part

266
00:15:32,560 --> 00:15:36,320
Speaker 5: what makes it so beneficial that it stays in alignment

267
00:15:36,400 --> 00:15:39,880
Speaker 5: with the objectives that you set for the technology over time.

268
00:15:40,520 --> 00:15:44,360
Speaker 5: So you can't do that without a robust governance process

269
00:15:44,360 --> 00:15:47,840
Speaker 5: in place. So we work with clients to share our

270
00:15:47,880 --> 00:15:50,360
Speaker 5: own story here at IBM in terms of how we

271
00:15:50,400 --> 00:15:54,000
Speaker 5: put that in place, but also in our consulting practice

272
00:15:54,760 --> 00:15:59,240
Speaker 5: to help clients work with these new generative capabilities and

273
00:15:59,280 --> 00:16:02,240
Speaker 5: foundation model and the like in order to put them

274
00:16:02,240 --> 00:16:04,280
Speaker 5: to work for their business in a way that's going

275
00:16:04,320 --> 00:16:06,960
Speaker 5: to be impactful to that business, but at the same

276
00:16:07,040 --> 00:16:08,080
Speaker 5: time be trusted.

277
00:16:08,280 --> 00:16:10,120
Speaker 4: So now I wanted to turn a little bit towards

278
00:16:10,120 --> 00:16:13,400
Speaker 4: Watson X governance, and so IBM recently announced their AI

279
00:16:13,480 --> 00:16:17,640
Speaker 4: platform Watson X, which will include a governance component. Could

280
00:16:17,680 --> 00:16:20,360
Speaker 4: you tell us a little more about watsonx dot governance.

281
00:16:21,000 --> 00:16:23,240
Speaker 5: Yeah, I mean before I do that, I'll just back

282
00:16:23,320 --> 00:16:26,920
Speaker 5: up and talk about the full platform and then lean

283
00:16:26,960 --> 00:16:29,840
Speaker 5: into Watson X because I think it's important to understand

284
00:16:30,120 --> 00:16:36,040
Speaker 5: the delivery of a full suite of capabilities, to get data,

285
00:16:36,120 --> 00:16:39,040
Speaker 5: to train models, and then to govern them over their

286
00:16:39,080 --> 00:16:44,600
Speaker 5: life cycle. All of these things are really important. From

287
00:16:44,640 --> 00:16:47,480
Speaker 5: the onset, you need to make sure that you have,

288
00:16:48,120 --> 00:16:52,360
Speaker 5: you know, for our WATSONX dot AI for example, that's

289
00:16:52,400 --> 00:16:56,960
Speaker 5: the studio to train new foundation models and generative AI

290
00:16:57,040 --> 00:17:02,680
Speaker 5: and machine learning capabilities, and we are populating that studio

291
00:17:03,040 --> 00:17:08,320
Speaker 5: with some IBM trained foundation models, which we're curating and

292
00:17:08,400 --> 00:17:12,159
Speaker 5: tailoring more specifically for enterprises. So that's really important. It

293
00:17:12,160 --> 00:17:14,520
Speaker 5: comes back to the point I made earlier about business

294
00:17:14,560 --> 00:17:21,240
Speaker 5: trust and the need to have enterprise ready technologies in

295
00:17:21,280 --> 00:17:25,439
Speaker 5: the AI space, and then the watsonex dot data is

296
00:17:25,480 --> 00:17:28,160
Speaker 5: a fit for purpose data store or a data lake,

297
00:17:28,760 --> 00:17:33,240
Speaker 5: and then watsonex dot gov. So that's a particular component

298
00:17:33,400 --> 00:17:37,360
Speaker 5: of the platform that my team and the AI Ethics

299
00:17:37,400 --> 00:17:41,440
Speaker 5: Board has really worked closely with the product team on developing,

300
00:17:41,520 --> 00:17:44,360
Speaker 5: and we're using it internally here in the Chief Privacy

301
00:17:44,400 --> 00:17:48,720
Speaker 5: Office as well to help us govern our own uses

302
00:17:48,840 --> 00:17:53,800
Speaker 5: of AI technology and our compliance program here. And it

303
00:17:54,000 --> 00:17:59,320
Speaker 5: essentially helps to notify you if a model becomes biased

304
00:17:59,400 --> 00:18:02,560
Speaker 5: or get out of alignment as you're using it over time.

305
00:18:03,119 --> 00:18:05,760
Speaker 5: So companies are going to need these capabilities. I mean

306
00:18:05,760 --> 00:18:09,960
Speaker 5: they need them today to deliver technologies with trust. They'll

307
00:18:10,000 --> 00:18:13,280
Speaker 5: need them tomorrow to comply with regulation which is on

308
00:18:13,320 --> 00:18:14,400
Speaker 5: the horizon.

309
00:18:14,000 --> 00:18:16,800
Speaker 4: And I think compliance becomes even more complex when you

310
00:18:16,840 --> 00:18:20,960
Speaker 4: consider international data protection laws and regulations. Honestly, I don't

311
00:18:21,000 --> 00:18:23,560
Speaker 4: know how anyone on any company legal team is keeping

312
00:18:23,640 --> 00:18:25,879
Speaker 4: up with this these days. But my question for you

313
00:18:26,000 --> 00:18:29,600
Speaker 4: is really how can businesses develop a strategy to maintain

314
00:18:29,640 --> 00:18:32,640
Speaker 4: compliance and to deal with it in this ever changing landscape.

315
00:18:32,760 --> 00:18:36,720
Speaker 5: It's increasingly more challenging. In fact, I saw statistic just

316
00:18:36,800 --> 00:18:41,360
Speaker 5: this morning that the regulatory obligations on companies have increased

317
00:18:41,400 --> 00:18:45,080
Speaker 5: something like seven hundred times in the last twenty years.

318
00:18:45,080 --> 00:18:49,679
Speaker 5: So it really is a huge focus area for companies.

319
00:18:49,840 --> 00:18:52,720
Speaker 5: You have to have a process in place in order

320
00:18:52,760 --> 00:18:55,320
Speaker 5: to do that, and it's not easy, particularly for a

321
00:18:55,359 --> 00:18:59,200
Speaker 5: company like IBM that it has a presence in over

322
00:18:59,240 --> 00:19:02,280
Speaker 5: one hundred and seven countries around the world. There is

323
00:19:02,320 --> 00:19:07,160
Speaker 5: more than one hundred and fifty comprehensive privacy regulations, there

324
00:19:07,200 --> 00:19:11,760
Speaker 5: are regulations of non personal data, there are AI regulations emerging,

325
00:19:12,720 --> 00:19:16,800
Speaker 5: so you really need an operational approach to it in

326
00:19:16,920 --> 00:19:18,919
Speaker 5: order to stay compliant. But one of the things we

327
00:19:18,960 --> 00:19:21,159
Speaker 5: do is we set a baseline. And a lot of

328
00:19:21,160 --> 00:19:24,600
Speaker 5: companies do this as well. So we define a privacy baseline,

329
00:19:24,640 --> 00:19:29,439
Speaker 5: we define an AI baseline, and we ensure then as

330
00:19:29,480 --> 00:19:31,880
Speaker 5: a result of that that there are very few deviances

331
00:19:31,920 --> 00:19:35,040
Speaker 5: because it incorporates in that baseline. So that's one of

332
00:19:35,080 --> 00:19:37,199
Speaker 5: the ways we do it. Other companies I think are

333
00:19:37,320 --> 00:19:42,520
Speaker 5: similarly situated in terms of doing that. But again, it

334
00:19:42,680 --> 00:19:45,359
Speaker 5: is a real challenge for global companies. It's one of

335
00:19:45,359 --> 00:19:49,320
Speaker 5: the reasons why we advocate for as much alignment as

336
00:19:49,400 --> 00:19:54,760
Speaker 5: possible on the international realm as well as nationally here

337
00:19:54,800 --> 00:19:58,560
Speaker 5: in the US, as much alignment as possible to make

338
00:19:58,800 --> 00:20:03,760
Speaker 5: compliance easier for easier and not just because companies want

339
00:20:03,800 --> 00:20:06,919
Speaker 5: an easy way to comply. But the harder it is,

340
00:20:07,200 --> 00:20:11,080
Speaker 5: the less likely there will be compliance. And it's not

341
00:20:11,200 --> 00:20:17,280
Speaker 5: the objective of anybody, governments, companies, consumers to have to

342
00:20:17,440 --> 00:20:20,920
Speaker 5: set legal obligations that companies simply can't meet.

343
00:20:21,280 --> 00:20:23,280
Speaker 4: And so what advice would you give to other companies

344
00:20:23,320 --> 00:20:26,040
Speaker 4: who are looking to rethink or strengthen their approach to AI.

345
00:20:26,080 --> 00:20:28,280
Speaker 5: Government think you need to start with, as we did,

346
00:20:28,400 --> 00:20:32,920
Speaker 5: foundational principles, and you need to start making decisions about

347
00:20:32,960 --> 00:20:36,000
Speaker 5: what technology you're going to deploy and what technology you're not,

348
00:20:36,080 --> 00:20:37,040
Speaker 5: what are you going to use it for, and what

349
00:20:37,040 --> 00:20:38,560
Speaker 5: aren't you going to use it for, and then when

350
00:20:38,600 --> 00:20:43,080
Speaker 5: you do use it, align to those principles. That's really important.

351
00:20:43,119 --> 00:20:47,680
Speaker 5: Formalize a program, have someone within the organization, whether it's

352
00:20:47,680 --> 00:20:52,000
Speaker 5: the chief Privacy officer, whether it's some other role, a

353
00:20:52,080 --> 00:20:56,520
Speaker 5: chief AI ethics officer, but have an accountable individual and

354
00:20:56,560 --> 00:21:00,760
Speaker 5: accountable organization. Do a maturity assessment, figure out where you

355
00:21:00,840 --> 00:21:03,560
Speaker 5: are and where you need to be, and really start

356
00:21:03,800 --> 00:21:07,960
Speaker 5: you know, putting it into place today. Don't wait for

357
00:21:08,320 --> 00:21:11,399
Speaker 5: regulation to apply directly to your business because it'll be

358
00:21:11,440 --> 00:21:11,880
Speaker 5: too late.

359
00:21:12,840 --> 00:21:16,240
Speaker 4: So Smart Talks features new creators, these visionaries like yourself

360
00:21:16,240 --> 00:21:19,800
Speaker 4: who are creatively applying technology in business to drive change.

361
00:21:20,000 --> 00:21:22,720
Speaker 4: I'm curious if you see yourself as creative.

362
00:21:23,160 --> 00:21:26,439
Speaker 5: You know, I definitely do. I mean, you need to

363
00:21:26,560 --> 00:21:31,200
Speaker 5: be creative when you're working in an industry that evolves

364
00:21:31,240 --> 00:21:35,720
Speaker 5: so very quickly. So you know, I started with IBM

365
00:21:36,000 --> 00:21:39,119
Speaker 5: when we were primarily a hardware company, right and we've

366
00:21:39,200 --> 00:21:42,600
Speaker 5: changed our business so significantly over the years. And the

367
00:21:42,720 --> 00:21:47,160
Speaker 5: issues that are raised with respect to each new technology,

368
00:21:47,200 --> 00:21:50,840
Speaker 5: whether it be cloud, whether it be AI now where

369
00:21:50,840 --> 00:21:52,439
Speaker 5: we're seeing a ton of issues, or you look at

370
00:21:52,440 --> 00:21:56,560
Speaker 5: emergent issues in the space of things like neurotechnologies and

371
00:21:56,680 --> 00:22:03,280
Speaker 5: quantum computers. You have to be strategic and you have

372
00:22:03,400 --> 00:22:06,640
Speaker 5: to be creative and thinking about how you can adapt

373
00:22:07,280 --> 00:22:12,520
Speaker 5: agilely quickly a company to an environment that is changing

374
00:22:12,520 --> 00:22:14,280
Speaker 5: so quickly and.

375
00:22:14,280 --> 00:22:17,359
Speaker 4: With this transformation happening at such a rapid pace. Do

376
00:22:17,400 --> 00:22:19,440
Speaker 4: you think creativity plays a role in how you think

377
00:22:19,480 --> 00:22:22,800
Speaker 4: about and implement, specifically a trustworthy AI strategy.

378
00:22:25,280 --> 00:22:29,320
Speaker 5: Yeah, I absolutely think it does. Because again it comes

379
00:22:29,320 --> 00:22:32,520
Speaker 5: back to these capabilities, and there are ways. I guess

380
00:22:32,720 --> 00:22:36,440
Speaker 5: how you define creativity could be different, right, But I'm

381
00:22:36,480 --> 00:22:39,720
Speaker 5: thinking of creativity in the sense of sort of agility

382
00:22:39,760 --> 00:22:43,840
Speaker 5: and strategic vision and creative problem solving. I think that's

383
00:22:44,119 --> 00:22:47,200
Speaker 5: really important in the world that we're in right now,

384
00:22:47,240 --> 00:22:51,840
Speaker 5: being able to creatively problem solve with new issues that

385
00:22:51,920 --> 00:22:55,000
Speaker 5: are rising sort of every day.

386
00:22:55,400 --> 00:22:56,920
Speaker 4: And so, how do you see the role of chief

387
00:22:56,960 --> 00:23:00,600
Speaker 4: privacy officer evolving in the future as AI technology continues

388
00:23:00,640 --> 00:23:03,520
Speaker 4: to advance, Like what stuff should CPOs take to stay

389
00:23:03,520 --> 00:23:05,440
Speaker 4: ahead of all these changes that are come in their way?

390
00:23:07,000 --> 00:23:10,920
Speaker 5: So the role is evolving in most companies, I would

391
00:23:10,960 --> 00:23:15,920
Speaker 5: say pretty rapidly. Many companies are looking to chief privacy

392
00:23:15,960 --> 00:23:19,480
Speaker 5: officers who are ready understand the data that's being used

393
00:23:19,480 --> 00:23:22,960
Speaker 5: in the organization and have programs to ensure compliance with

394
00:23:23,119 --> 00:23:26,920
Speaker 5: laws that require you to manage that data in accordance

395
00:23:26,920 --> 00:23:29,520
Speaker 5: with data protection laws and the like. It's a natural

396
00:23:29,600 --> 00:23:35,560
Speaker 5: place and position for AI responsibility. And so I think

397
00:23:35,600 --> 00:23:38,240
Speaker 5: what's happening to a lot of chief privacy officers is

398
00:23:38,280 --> 00:23:41,800
Speaker 5: they're being asked to take on this AI governance responsibility

399
00:23:41,840 --> 00:23:45,320
Speaker 5: for companies and if not take it on at least

400
00:23:45,359 --> 00:23:48,800
Speaker 5: play a very key role working with other parts of

401
00:23:48,840 --> 00:23:52,479
Speaker 5: the business in AI governance. So that really is changing.

402
00:23:52,720 --> 00:23:56,760
Speaker 5: And if chief privacy officers are in companies who maybe

403
00:23:56,800 --> 00:24:00,560
Speaker 5: haven't started thinking about AI yet, they should, So I

404
00:24:00,560 --> 00:24:04,439
Speaker 5: would encourage them to look at different resources that are

405
00:24:04,480 --> 00:24:08,840
Speaker 5: available already in AI governance space. For example, the International

406
00:24:08,880 --> 00:24:13,320
Speaker 5: Association of Privacy Professionals, which is the seventy five thousand

407
00:24:13,359 --> 00:24:18,440
Speaker 5: member professional body for the profession of Chief Privacy officers,

408
00:24:18,560 --> 00:24:23,200
Speaker 5: just recently launched an AI Governance Initiative and an AI

409
00:24:23,240 --> 00:24:27,199
Speaker 5: Governance certification program. I sit on their advisory board. But

410
00:24:27,320 --> 00:24:30,000
Speaker 5: that's just emblematic of the fact that the field is

411
00:24:30,080 --> 00:24:31,520
Speaker 5: changing so rapidly.

412
00:24:32,760 --> 00:24:35,320
Speaker 4: And so, you know, speaking of rapid change, when you're

413
00:24:35,600 --> 00:24:38,000
Speaker 4: back here on smart Talks in twenty twenty one, you

414
00:24:38,040 --> 00:24:40,359
Speaker 4: said that the future of AI will be more transparent

415
00:24:40,400 --> 00:24:42,639
Speaker 4: and more trustworthy. You know, what do you see the

416
00:24:42,640 --> 00:24:44,679
Speaker 4: next five to ten years holding? You know, when you're

417
00:24:44,720 --> 00:24:47,639
Speaker 4: back on smart Talks in you know, twenty twenty six,

418
00:24:47,720 --> 00:24:49,399
Speaker 4: you know twenty thirty, you know what are we going

419
00:24:49,480 --> 00:24:51,760
Speaker 4: to be talking about when it comes to AI technology

420
00:24:51,760 --> 00:24:52,400
Speaker 4: and governance.

421
00:24:53,240 --> 00:24:55,359
Speaker 5: So I try to be an optimist, right and I

422
00:24:55,440 --> 00:24:59,280
Speaker 5: said that two years ago, and I think we're seeing

423
00:24:59,320 --> 00:25:03,720
Speaker 5: it now come into fruition, and there will be requirements,

424
00:25:04,520 --> 00:25:07,320
Speaker 5: whether they're coming from the US, whether they're coming from Europe,

425
00:25:07,400 --> 00:25:11,000
Speaker 5: whether they're just coming from voluntary adoption by clients of

426
00:25:11,080 --> 00:25:15,800
Speaker 5: things like the NIST Risk Management Framework, really important voluntary frameworks.

427
00:25:16,560 --> 00:25:20,160
Speaker 5: You're going to have to adopt transparent and explainable practices

428
00:25:20,320 --> 00:25:22,919
Speaker 5: in your uses of AI. So I do see that happening.

429
00:25:23,000 --> 00:25:25,439
Speaker 5: And in the next five to ten years, boy, I

430
00:25:25,440 --> 00:25:31,439
Speaker 5: think we'll see more research into trust in techniques because

431
00:25:31,440 --> 00:25:35,399
Speaker 5: we don't really know, for example, how to water mark.

432
00:25:36,160 --> 00:25:39,680
Speaker 5: We're calling for things like watermarking. There'll be more research

433
00:25:39,720 --> 00:25:44,800
Speaker 5: into how to do that. I think you'll see you

434
00:25:45,040 --> 00:25:48,000
Speaker 5: regulation that's specifically going to require those types of things.

435
00:25:48,040 --> 00:25:50,199
Speaker 5: So I think again, I think the regulation is going

436
00:25:50,240 --> 00:25:53,040
Speaker 5: to drive research. It's going to drive research into these

437
00:25:53,119 --> 00:25:58,280
Speaker 5: areas that will help ensure that we can deliver new capabilities,

438
00:25:58,320 --> 00:26:01,680
Speaker 5: generative capabilities, and the life with trust and explainability.

439
00:26:01,880 --> 00:26:03,920
Speaker 4: Thank you so much, Christina for joining me on smart

440
00:26:03,920 --> 00:26:05,600
Speaker 4: Talks to talk about AI and governance.

441
00:26:06,560 --> 00:26:08,560
Speaker 5: Well, thank you very much for having me.

442
00:26:09,840 --> 00:26:14,560
Speaker 3: To unlock the transformative growth possible with artificial intelligence, businesses

443
00:26:14,640 --> 00:26:17,760
Speaker 3: need to know what they wish to grow into first.

444
00:26:18,720 --> 00:26:21,600
Speaker 3: Like Christina said, the best way forward in the AI

445
00:26:21,680 --> 00:26:25,720
Speaker 3: future is for businesses to figure out their own foundational

446
00:26:25,760 --> 00:26:30,359
Speaker 3: principles around using the technology, drawing on those principles to

447
00:26:30,440 --> 00:26:34,240
Speaker 3: apply AI in a way that's ethically consistent with their

448
00:26:34,280 --> 00:26:37,960
Speaker 3: mission and complies with the legal frameworks built to hold

449
00:26:38,000 --> 00:26:42,960
Speaker 3: the technology accountable. As AI adoption grows more and more widespread,

450
00:26:43,200 --> 00:26:47,560
Speaker 3: so too will the expectation from consumers and regulators that

451
00:26:47,600 --> 00:26:53,040
Speaker 3: businesses use it responsibly. Investing independable AI governance is a

452
00:26:53,080 --> 00:26:57,040
Speaker 3: way for businesses to lay the foundations for technology that

453
00:26:57,080 --> 00:27:01,000
Speaker 3: their customers can trust while rising to the challenge of

454
00:27:01,080 --> 00:27:06,879
Speaker 3: increasing regulatory complexity. Though the emergence of AI does complicate

455
00:27:07,000 --> 00:27:11,560
Speaker 3: an already tough compliance landscape, businesses now face a creative

456
00:27:11,600 --> 00:27:16,240
Speaker 3: opportunity to set a precedent for what accountability in AI

457
00:27:16,400 --> 00:27:20,359
Speaker 3: looks like and rethink what it means to deploy trustworthy

458
00:27:20,800 --> 00:27:26,640
Speaker 3: artificial intelligence. I'm Malcolm Gladwell. This is a paid advertisement

459
00:27:27,040 --> 00:27:30,320
Speaker 3: from IBM. Smart Talks with IBM will be taking a

460
00:27:30,359 --> 00:27:34,480
Speaker 3: short hiatus, but look for new episodes in the coming weeks.

461
00:27:35,160 --> 00:27:38,560
Speaker 3: Smart Talks with IBM is produced by Matt Ramano, David

462
00:27:38,640 --> 00:27:43,680
Speaker 3: jaw nische Venkat and Royston Deserve with Jacob Goldstein. We're

463
00:27:43,800 --> 00:27:47,480
Speaker 3: edited by Lydia gene Kott. Our engineer is Jason Gambrel.

464
00:27:47,800 --> 00:27:52,720
Speaker 3: Theme song by Gramoscope. Special thanks to Carli Migliori, Andy Kelly,

465
00:27:53,119 --> 00:27:57,400
Speaker 3: Kathy Callahan and the eight Bar and IBM teams, as

466
00:27:57,400 --> 00:28:01,280
Speaker 3: well as the Pushkin marketing team. Smart Talks with ib

467
00:28:01,440 --> 00:28:04,800
Speaker 3: M is a production of Pushkin Industries and Ruby Studio

468
00:28:05,200 --> 00:28:09,720
Speaker 3: at iHeartMedia. To find more Pushkin podcasts, listen on the

469
00:28:09,800 --> 00:28:23,920
Speaker 3: iHeartRadio app, Apple Podcasts, or wherever you listen to podcasts,