WEBVTT

1
00:00:00.125 --> 00:00:03.378
I'm Manveen Rana here in Davos,
and I'm joined by Jessica, who's

2
00:00:03.378 --> 00:00:06.589
been looking at AI, which is one
of the biggest themes here this

3
00:00:06.589 --> 00:00:06.840
year.

4
00:00:07.674 --> 00:00:11.928
Tell us, in particular, what is
the thinking in terms of what AI

5
00:00:11.928 --> 00:00:14.014
might mean for gender diversity?

6
00:00:14.514 --> 00:00:17.934
Yeah, there's been a lot of talk
about the representation of

7
00:00:17.934 --> 00:00:20.687
women in people that are
developing, coding, and

8
00:00:20.687 --> 00:00:21.479
regulating AI.

9
00:00:21.479 --> 00:00:25.525
I actually think there's a whole
other impact of AI that is not

10
00:00:25.525 --> 00:00:29.112
talked about enough, which is
what does it do to female

11
00:00:29.112 --> 00:00:30.030
representation?

12
00:00:30.363 --> 00:00:34.117
If you think of your virtual
experience, all the avatars, all

13
00:00:34.117 --> 00:00:37.704
of the voices you're interacting
with, they're all female.

14
00:00:38.121 --> 00:00:39.539
They're overarchingly female.

15
00:00:39.748 --> 00:00:42.250
And when you do user tests, that
is what people want.

16
00:00:42.250 --> 00:00:46.421
They actually want to interact
with female agents in a massive

17
00:00:46.421 --> 00:00:48.923
way, digitally, and at the same
time,

18
00:00:48.923 --> 00:00:51.718
there's a very negative effect
of that from a diversity

19
00:00:51.718 --> 00:00:52.260
standpoint.

20
00:00:52.469 --> 00:00:56.347
Basically, everybody's ordering
females around all the time.

21
00:00:57.432 --> 00:01:00.685
How will that impact the way
people think about the role of

22
00:01:00.685 --> 00:01:01.686
females in society?

23
00:01:02.103 --> 00:01:05.190
That is a theme that is totally
absent from the discussion, and

24
00:01:05.190 --> 00:01:06.900
I wish it was much more
prevalent.

25
00:01:06.941 --> 00:01:07.776
You're right,

26
00:01:07.776 --> 00:01:10.028
I don't think people think about
it, but it's Alexa this, it's

27
00:01:10.028 --> 00:01:10.403
Alexa that.

28
00:01:10.653 --> 00:01:11.780
What do you want to sort of see?

29
00:01:11.780 --> 00:01:13.615
Should there be more male voices

30
00:01:13.615 --> 00:01:14.449
you're asking for help?

31
00:01:15.033 --> 00:01:16.242
Well, there's two ways about
that.

32
00:01:16.242 --> 00:01:19.746
You can do more male voices or
you can try to aim for more

33
00:01:19.746 --> 00:01:23.124
robotic voices, and it's
actually a real trust question.

34
00:01:23.124 --> 00:01:27.045
Do you want the voices that are
there and the embodiment of the

35
00:01:27.045 --> 00:01:30.757
avatars that are used to look
human-like, or to be what they

36
00:01:30.757 --> 00:01:34.344
are: robots, so that you can
identify when you're actually

37
00:01:34.344 --> 00:01:36.304
interacting with a human or not?

38
00:01:36.846 --> 00:01:39.641
That is the type of question I
think corporates should be

39
00:01:39.641 --> 00:01:40.517
asking themselves.

40
00:01:40.809 --> 00:01:44.687
They're part of what I call the
micro-ethical decisions of AI.

41
00:01:45.146 --> 00:01:48.817
And today, the reality in most
corporations, they're taken in a

42
00:01:48.817 --> 00:01:52.153
very decentralized fashion,
whereas I think they should be

43
00:01:52.153 --> 00:01:54.155
central in the corporate
strategy.

44
00:01:54.572 --> 00:01:56.616
Jessica, that's absolutely
fascinating.

45
00:01:56.616 --> 00:01:58.118
Thanks so much for joining us
here.

46
00:01:58.243 --> 00:01:58.827
My pleasure.

47
00:01:58.827 --> 00:01:59.702
Thank you for having me.