1
00:00:02,040 --> 00:00:07,040
Speaker 1: Welcome to brain Stuff from How Stuff Works, Hey, brain Stuff,

2
00:00:07,080 --> 00:00:12,960
Speaker 1: Lauren Bogelbaum here. The date is January nine nine. Robert Williams,

3
00:00:13,000 --> 00:00:15,960
Speaker 1: a year old factory worker in a Ford Motor Company

4
00:00:16,000 --> 00:00:19,040
Speaker 1: casting plant in Flat Rock, Michigan, is asked to scale

5
00:00:19,079 --> 00:00:22,160
Speaker 1: a massive shelving unit to manually count the parts there.

6
00:00:22,560 --> 00:00:25,520
Speaker 1: The five story machine used to retrieve the castings is

7
00:00:25,560 --> 00:00:28,800
Speaker 1: giving false readings, and it's William's task to go up

8
00:00:28,840 --> 00:00:32,400
Speaker 1: and find out how many there actually are. While Williams

9
00:00:32,440 --> 00:00:35,000
Speaker 1: is up there doing the job, a robot arm, also

10
00:00:35,040 --> 00:00:38,400
Speaker 1: tasked with parts retrieval, goes about its work. In doing so,

11
00:00:38,520 --> 00:00:41,960
Speaker 1: their paths accidentally cross, and the robotic arm catches the

12
00:00:42,000 --> 00:00:44,839
Speaker 1: young man, striking him in the head and killing him instantly,

13
00:00:45,320 --> 00:00:48,440
Speaker 1: And thus, on this wintry day, Williams becomes the first

14
00:00:48,560 --> 00:00:52,200
Speaker 1: human in history to be killed by a robot. The death,

15
00:00:52,240 --> 00:00:55,880
Speaker 1: of course, was completely unintentional. There simply weren't safeguards in

16
00:00:55,920 --> 00:00:59,000
Speaker 1: place to protect Williams. No alarms notified him of the

17
00:00:59,000 --> 00:01:01,440
Speaker 1: approaching arm, and there was no technology in place to

18
00:01:01,520 --> 00:01:04,679
Speaker 1: alter the robot's behavior in the presence of a human person.

19
00:01:05,360 --> 00:01:08,600
Speaker 1: As of nineteen seventy nine, the artificial intelligence involved wasn't

20
00:01:08,600 --> 00:01:11,520
Speaker 1: sophisticated enough to do anything to prevent such a death.

21
00:01:12,160 --> 00:01:14,640
Speaker 1: Jury agreed that not enough care had been put into

22
00:01:14,680 --> 00:01:17,640
Speaker 1: the design of the robot. William's family wanted ten million

23
00:01:17,680 --> 00:01:20,800
Speaker 1: dollar lawsuit for his wrongful death from the Unit Handling Systems,

24
00:01:21,000 --> 00:01:24,400
Speaker 1: a division of Litton Industries, the manufacturer that designed the robot.

25
00:01:25,120 --> 00:01:29,080
Speaker 1: In the ensuing years, roboticists, computer scientists, and artificial intelligence

26
00:01:29,120 --> 00:01:31,720
Speaker 1: experts have continued to struggle with the issue of how

27
00:01:31,800 --> 00:01:35,839
Speaker 1: robots can safely interact with humans without causing them harm.

28
00:01:35,959 --> 00:01:39,000
Speaker 1: Decades later, reports of human deaths caused by robots or

29
00:01:39,120 --> 00:01:43,440
Speaker 1: artificial intelligences feel more commonplace. Uber and Tesla have made

30
00:01:43,440 --> 00:01:46,360
Speaker 1: the news with reports of their partially autonomous cars getting

31
00:01:46,360 --> 00:01:50,640
Speaker 1: into accidents and killing passengers or striking pedestrians. Though many

32
00:01:50,720 --> 00:01:55,240
Speaker 1: safeguards now are in place, the problem still hasn't been solved. However,

33
00:01:55,320 --> 00:01:57,440
Speaker 1: none of these deaths are caused by the will of

34
00:01:57,480 --> 00:02:00,640
Speaker 1: the robot. These programmed machines don't to have a will,

35
00:02:00,760 --> 00:02:03,720
Speaker 1: just a set of operating conditions. But there's a worry

36
00:02:03,800 --> 00:02:06,280
Speaker 1: fanned by the flames of science fiction stories like The

37
00:02:06,360 --> 00:02:09,720
Speaker 1: Terminator or the Matrix, that artificial intelligences could develop a

38
00:02:09,720 --> 00:02:13,000
Speaker 1: will of their own, and in that development, the desire

39
00:02:13,040 --> 00:02:16,560
Speaker 1: to harm a human. We spoke with Shaman Whitson, Associate

40
00:02:16,560 --> 00:02:19,200
Speaker 1: professor in the Department of Computer Science at the University

41
00:02:19,200 --> 00:02:22,560
Speaker 1: of Oxford and chief scientist and co founder of Morphy's Labs.

42
00:02:22,919 --> 00:02:26,800
Speaker 1: He calls this concern the anthropomorphic fallacy. This, he said,

43
00:02:26,960 --> 00:02:30,000
Speaker 1: is the assumption that a system with human like intelligence

44
00:02:30,080 --> 00:02:34,120
Speaker 1: must also have humanlike desires e g. To survive, be free,

45
00:02:34,200 --> 00:02:37,600
Speaker 1: have dignity, et cetera. There's absolutely no reason why this

46
00:02:37,639 --> 00:02:39,840
Speaker 1: would be the case, as such a system will only

47
00:02:39,880 --> 00:02:44,320
Speaker 1: have whatever desires we give it. Value misalignment, he argues,

48
00:02:44,440 --> 00:02:47,799
Speaker 1: is the greater existential threat, where a gap exists between

49
00:02:47,800 --> 00:02:50,000
Speaker 1: what a programmer tells a machine to do and what

50
00:02:50,080 --> 00:02:53,960
Speaker 1: the programmer really meant to happen. Whitson explained, how do

51
00:02:54,000 --> 00:02:56,960
Speaker 1: you communicate your values to an intelligent system such that

52
00:02:57,000 --> 00:03:00,640
Speaker 1: the actions it takes fulfill your true intentions. The discrepancy

53
00:03:00,680 --> 00:03:03,639
Speaker 1: between the two becomes more consequential as the computer becomes

54
00:03:03,639 --> 00:03:07,640
Speaker 1: more intelligent and autonomous. Whiteston tells us that the even

55
00:03:07,680 --> 00:03:11,320
Speaker 1: greater threat, however, is scientists purposefully designing robots that can

56
00:03:11,400 --> 00:03:15,440
Speaker 1: kill human targets without human intervention for military purposes. The

57
00:03:15,440 --> 00:03:18,680
Speaker 1: Guardian reports that three d and eighty one partly autonomous

58
00:03:18,720 --> 00:03:22,079
Speaker 1: weapon and military robotics systems have been deployed or are

59
00:03:22,160 --> 00:03:25,800
Speaker 1: under development in a dozen countries, including China, France, Israel,

60
00:03:26,000 --> 00:03:31,280
Speaker 1: the UK and the US. In AI and robotics researchers

61
00:03:31,320 --> 00:03:33,680
Speaker 1: around the world published an open letter calling for a

62
00:03:33,720 --> 00:03:37,360
Speaker 1: worldwide ban on such technology, and it's currently endorsed by

63
00:03:37,360 --> 00:03:40,200
Speaker 1: almost four thousand researchers in the field and over twenty

64
00:03:40,240 --> 00:03:43,840
Speaker 1: two thousand other concerned humans, including folks like Stephen Hawking

65
00:03:43,880 --> 00:03:46,839
Speaker 1: and Noam Chomsky, and the United Nations is meeting again

66
00:03:46,880 --> 00:03:50,680
Speaker 1: in to discuss if and how to regulate so called

67
00:03:50,800 --> 00:03:58,840
Speaker 1: killer robots. Today's episode was written by Brian Young and

68
00:03:58,880 --> 00:04:01,480
Speaker 1: produced by Tyler Klang. For more in this and lots

69
00:04:01,480 --> 00:04:04,080
Speaker 1: of other all two human topics, visit our home planet,

70
00:04:04,280 --> 00:04:16,839
Speaker 1: TAW staff works dot com.