1
00:00:03,400 --> 00:00:06,440
Speaker 1: You're listening to Disinformed, a mini series from There Are

2
00:00:06,440 --> 00:00:13,360
Speaker 1: No Girls on the Internet. I'm Bridget Todd. I talk

3
00:00:13,440 --> 00:00:17,079
Speaker 1: a lot about hostility towards underrepresented voices and technology and

4
00:00:17,239 --> 00:00:20,000
Speaker 1: how it ultimately hurts us all, and facial recognition is

5
00:00:20,000 --> 00:00:22,279
Speaker 1: a great example of what I mean. We know that

6
00:00:22,280 --> 00:00:25,200
Speaker 1: facial recognition technology is unreliable when it comes to people

7
00:00:25,200 --> 00:00:28,040
Speaker 1: with darker skin tones and women's faces, and with the

8
00:00:28,120 --> 00:00:30,560
Speaker 1: rise of this technology being used for things like policing

9
00:00:30,560 --> 00:00:33,800
Speaker 1: and surveillance, this is obviously a pretty big problem. We've

10
00:00:33,840 --> 00:00:36,280
Speaker 1: already seen reports of black people being arrested for crimes

11
00:00:36,360 --> 00:00:38,600
Speaker 1: they had nothing to do with, based solely on the

12
00:00:38,680 --> 00:00:41,800
Speaker 1: use of faulty facial recognition software. This is what happens

13
00:00:41,800 --> 00:00:44,519
Speaker 1: when black women are treated like outsiders in tech spaces.

14
00:00:45,040 --> 00:00:47,000
Speaker 1: If the technology that plays such a big role in

15
00:00:47,040 --> 00:00:49,120
Speaker 1: all of our lives is built by teams of homogeneous

16
00:00:49,159 --> 00:00:52,720
Speaker 1: people with biases, they might not even realize that technology

17
00:00:52,760 --> 00:00:55,440
Speaker 1: can go on to harm us all. Now, the reason

18
00:00:55,480 --> 00:00:57,880
Speaker 1: that we know about the gender and racial biases encoded

19
00:00:57,880 --> 00:01:00,680
Speaker 1: into facial recognition technology at all is because of the

20
00:01:00,760 --> 00:01:04,160
Speaker 1: labor and talent of black women AI researchers who studied it.

21
00:01:04,800 --> 00:01:08,320
Speaker 1: Joy Boulaweeny, Tim Net Gabaru, and deb Raji are black

22
00:01:08,319 --> 00:01:11,880
Speaker 1: female scientists who have conducted groundbreaking research under racial and

23
00:01:11,920 --> 00:01:15,720
Speaker 1: gender biases of facial recognition, software and artificial intelligence. If

24
00:01:15,760 --> 00:01:18,160
Speaker 1: you've seen the documentary Coded Bias, then you've seen some

25
00:01:18,200 --> 00:01:20,680
Speaker 1: of their amazing work. It's actually on Netflix right now,

26
00:01:20,720 --> 00:01:23,479
Speaker 1: and I highly suggest you'll check it out. But if

27
00:01:23,520 --> 00:01:26,160
Speaker 1: you watched sixty Minutes last week, you would never know

28
00:01:26,240 --> 00:01:29,360
Speaker 1: that it was black women who first did this groundbreaking research.

29
00:01:30,160 --> 00:01:33,119
Speaker 1: That's because even though Sixty Minutes producers reached out to Joy,

30
00:01:33,360 --> 00:01:35,640
Speaker 1: who is also the head of the Algorithmic Justice League,

31
00:01:35,720 --> 00:01:38,479
Speaker 1: an organization that raises awareness about the harms of AI,

32
00:01:38,560 --> 00:01:40,960
Speaker 1: they just erased her voice from a story about the

33
00:01:41,040 --> 00:01:43,640
Speaker 1: use of facial recognition to arrest to black people for

34
00:01:43,720 --> 00:01:47,720
Speaker 1: crimes they didn't commit. Instead, they interviewed Patrick Growther, a

35
00:01:47,720 --> 00:01:51,000
Speaker 1: white male computer scientist. Now, no one is saying that

36
00:01:51,040 --> 00:01:54,720
Speaker 1: Growther doesn't know his stuff, but he himself correctly cites

37
00:01:54,760 --> 00:01:57,480
Speaker 1: the work of these black women and acknowledges that there's

38
00:01:57,520 --> 00:02:00,840
Speaker 1: with the landmark study on bias and facial recognition, and

39
00:02:00,880 --> 00:02:03,000
Speaker 1: that their work was the motivation for his own work,

40
00:02:03,600 --> 00:02:05,840
Speaker 1: so why didn't CBS feel the need to cite them

41
00:02:05,880 --> 00:02:09,880
Speaker 1: at all? Even worse, Joyce says that CBS actually did

42
00:02:09,919 --> 00:02:12,520
Speaker 1: reach after her. She says that she spoke to sixty

43
00:02:12,560 --> 00:02:16,040
Speaker 1: minutes producers for hours and spent time building a custom

44
00:02:16,120 --> 00:02:19,400
Speaker 1: demo for Anderson Cooper and made recommendations on research to

45
00:02:19,440 --> 00:02:22,040
Speaker 1: include subjects to interview, with the emphasis on stories of

46
00:02:22,080 --> 00:02:24,440
Speaker 1: black people who have been falsely arrested because of a I.

47
00:02:25,040 --> 00:02:27,000
Speaker 1: She was on her way to get COVID tested for

48
00:02:27,200 --> 00:02:30,600
Speaker 1: the interview when at the very last minute it was canceled.

49
00:02:31,320 --> 00:02:34,680
Speaker 1: This is a ratio of black women's labor, voices, and talent,

50
00:02:35,639 --> 00:02:37,639
Speaker 1: and it's even worse than it happened on a segment

51
00:02:37,680 --> 00:02:41,320
Speaker 1: about racial bias. Why wouldn't CBS center the black woman

52
00:02:41,360 --> 00:02:44,120
Speaker 1: who pioneered this research on a story about the way

53
00:02:44,120 --> 00:02:46,880
Speaker 1: that it harms the black community. Why don't they only

54
00:02:46,919 --> 00:02:50,080
Speaker 1: talk to white researchers? And it's the same thing that

55
00:02:50,160 --> 00:02:52,680
Speaker 1: so many black women in tech have talked about. We

56
00:02:52,720 --> 00:02:56,040
Speaker 1: aren't seen as experts even in the fields that we build.

57
00:02:56,440 --> 00:02:59,959
Speaker 1: Joy is not taking us. She launched a petition demanding

58
00:03:00,080 --> 00:03:02,720
Speaker 1: CBS apologized and take specific steps to make sure that

59
00:03:02,760 --> 00:03:05,280
Speaker 1: black women are properly cited in the future, which just

60
00:03:05,320 --> 00:03:07,200
Speaker 1: goes to show you that black women are pretty much

61
00:03:07,200 --> 00:03:10,080
Speaker 1: always doing the work that makes things better for everyone.

62
00:03:10,560 --> 00:03:14,440
Speaker 1: Those reporting on algorithmic biases and discrimination must prioritize the

63
00:03:14,520 --> 00:03:17,400
Speaker 1: perspective of those doing the work, who are often marginalized

64
00:03:17,480 --> 00:03:22,200
Speaker 1: voices themselves. This eraser is routine, this is systemic. For centuries,

65
00:03:22,240 --> 00:03:24,560
Speaker 1: the labor of black women has been kept in the background,

66
00:03:24,760 --> 00:03:27,840
Speaker 1: behind closed doors, and out of the public view. Black

67
00:03:27,840 --> 00:03:30,880
Speaker 1: women and the brilliant, cutting, edge, field leading and sustaining

68
00:03:30,919 --> 00:03:33,639
Speaker 1: work that they do must not be erased any longer.

69
00:03:33,919 --> 00:03:36,320
Speaker 1: The petition reads and you can sign the petition at

70
00:03:36,320 --> 00:03:39,720
Speaker 1: the link in our show description. CBS and honestly, tech

71
00:03:39,760 --> 00:03:43,040
Speaker 1: in general needs to stop sidelining black women. It is

72
00:03:43,080 --> 00:03:45,840
Speaker 1: Black women who consistently do the work of making tech

73
00:03:45,920 --> 00:03:48,720
Speaker 1: safer and better for everyone, and we need to be

74
00:03:48,760 --> 00:03:58,360
Speaker 1: recognized for it, not erased or silenced. If you've enjoyed

75
00:03:58,360 --> 00:04:05,440
Speaker 1: this podcast, please help us grow by subscribing. Got a

76
00:04:05,480 --> 00:04:07,480
Speaker 1: story about an interesting thing in tech, or just want

77
00:04:07,520 --> 00:04:09,320
Speaker 1: to say hi. We'd love to hear from you at

78
00:04:09,320 --> 00:04:12,320
Speaker 1: Hello at tango dot com. Dis Informed is brought to

79
00:04:12,360 --> 00:04:14,280
Speaker 1: you by There are no girls on the Internet. It's

80
00:04:14,280 --> 00:04:17,760
Speaker 1: a production of iHeart Radio and Unbust Creative Jonathan Strickland

81
00:04:17,760 --> 00:04:21,000
Speaker 1: is our executive producer. Tory Harrison is our supervising producer

82
00:04:21,040 --> 00:04:24,920
Speaker 1: and engineer. Michaelmato is our contributing producer. I'm your host

83
00:04:24,960 --> 00:04:27,839
Speaker 1: Bridget Tod. For more great podcasts, check out the iHeart

84
00:04:27,920 --> 00:04:30,560
Speaker 1: Radio app, Apple podcast, or wherever you get your podcasts.