WEBVTT - Bots of War

0:00:03.640 --> 0:00:07.920
<v Speaker 1>In films like Terminator, autonomous weapons become an existential threat

0:00:07.960 --> 0:00:11.360
<v Speaker 1>to humanity. In reality, experts from around the world are

0:00:11.480 --> 0:00:14.680
<v Speaker 1>urging countries to back away from developing smart weapons. I'm

0:00:14.760 --> 0:00:20.880
<v Speaker 1>Jonathan Strickland, and this is text stuff. They it's a

0:00:21.000 --> 0:00:25.919
<v Speaker 1>terrifying thought. A weapon controlled by an artificially intelligent program

0:00:25.960 --> 0:00:30.639
<v Speaker 1>identifies targets on its own and unleasha's deadly violence against them.

0:00:30.680 --> 0:00:34.080
<v Speaker 1>There's no human guiding the machine. It's effectively a gun

0:00:34.360 --> 0:00:37.120
<v Speaker 1>that chooses whom it will shoot, all on its own.

0:00:37.760 --> 0:00:40.680
<v Speaker 1>It's not a far fetched idea. In an arrow where

0:00:40.680 --> 0:00:44.199
<v Speaker 1>computer programs can defeat humans at games like chess or go,

0:00:44.840 --> 0:00:47.360
<v Speaker 1>or one where cars are getting closer to taking over

0:00:47.400 --> 0:00:50.360
<v Speaker 1>the wheel and driving on their own, it's not difficult

0:00:50.360 --> 0:00:53.800
<v Speaker 1>to imagine a robot design specifically to perform the role

0:00:53.880 --> 0:00:57.880
<v Speaker 1>of soldiers. Militaries around the world already depend upon systems

0:00:57.880 --> 0:01:01.920
<v Speaker 1>that have semi autonomous capabilities, from tracking technologies to drones,

0:01:02.280 --> 0:01:05.440
<v Speaker 1>and many experts warned that consequences of developing and deploying

0:01:05.440 --> 0:01:09.520
<v Speaker 1>this technology are dire. One potential scenario is that such

0:01:09.560 --> 0:01:13.200
<v Speaker 1>devices could and likely will occasionally fire upon people who

0:01:13.280 --> 0:01:16.959
<v Speaker 1>aren't actually targets. During a gathering at the United Nations

0:01:16.959 --> 0:01:21.880
<v Speaker 1>on November four, a group of concerned AI experts presented

0:01:21.920 --> 0:01:25.240
<v Speaker 1>a film that includes such a scenario. In the movie,

0:01:25.440 --> 0:01:27.959
<v Speaker 1>a fleet of drones flies down and fires upon a

0:01:28.000 --> 0:01:32.040
<v Speaker 1>classroom filled with students. It's a horrifying thought and one

0:01:32.080 --> 0:01:35.440
<v Speaker 1>that could become reality should the development of this technology continue,

0:01:35.640 --> 0:01:39.840
<v Speaker 1>warrant these experts. Even if we assume everything always works

0:01:39.880 --> 0:01:44.280
<v Speaker 1>as intended, there are still huge ethical problems. For example,

0:01:44.560 --> 0:01:47.600
<v Speaker 1>if a nation has a sizeable army of robot soldiers,

0:01:47.800 --> 0:01:51.240
<v Speaker 1>whether in drone form or otherwise, would that nation be

0:01:51.320 --> 0:01:54.960
<v Speaker 1>more likely to enter into armed conflict with others? After all,

0:01:55.000 --> 0:01:57.720
<v Speaker 1>the stakes are lower for this country. Its soldiers are

0:01:57.760 --> 0:02:01.400
<v Speaker 1>made of metal and plastic, and while expensive, they can

0:02:01.400 --> 0:02:05.440
<v Speaker 1>be replaced. There are fewer dark consequences from the perspective

0:02:05.480 --> 0:02:09.320
<v Speaker 1>of loss with such an army. This isn't a new fear.

0:02:09.639 --> 0:02:12.440
<v Speaker 1>AI experts have been urging the United Nations to adopt

0:02:12.520 --> 0:02:15.520
<v Speaker 1>a ban on autonomous weapons for a few years now.

0:02:15.880 --> 0:02:19.960
<v Speaker 1>Back in when the first serious discussions began, many people

0:02:20.000 --> 0:02:22.880
<v Speaker 1>may have felt the whole conversation was premature, but as

0:02:22.919 --> 0:02:26.320
<v Speaker 1>the fields of machine intelligence, machine learning, and robotics have

0:02:26.360 --> 0:02:29.080
<v Speaker 1>all advanced over the years, it's apparent that this future

0:02:29.320 --> 0:02:31.800
<v Speaker 1>isn't as far off as we may have thought just

0:02:31.880 --> 0:02:35.000
<v Speaker 1>a few years ago. There are more than sixty non

0:02:35.120 --> 0:02:38.480
<v Speaker 1>governmental organizations that have banded together to call for an

0:02:38.480 --> 0:02:42.760
<v Speaker 1>outright ban on autonomous weapon development. Unfortunately, it's not as

0:02:42.800 --> 0:02:44.640
<v Speaker 1>simple as taking the floor at the u N and

0:02:44.720 --> 0:02:48.120
<v Speaker 1>convincing everyone that this technology poses far more threats than

0:02:48.160 --> 0:02:52.080
<v Speaker 1>benefits to humanity. For now, the forum for these discussions

0:02:52.120 --> 0:02:56.399
<v Speaker 1>is a subcommittee called the Convention on Certain Conventional Weapons,

0:02:56.680 --> 0:03:00.200
<v Speaker 1>a terrible name. It's a consensus based forum, and each

0:03:00.280 --> 0:03:03.320
<v Speaker 1>nation has the power to veto any proposed ban. As

0:03:03.320 --> 0:03:06.680
<v Speaker 1>you might imagine, it's pretty difficult to convince more than

0:03:06.720 --> 0:03:09.840
<v Speaker 1>a hundred twenty countries to all agree on a single topic.

0:03:10.560 --> 0:03:15.240
<v Speaker 1>Earlier in a group of entrepreneurs, researchers, computer scientists, and

0:03:15.320 --> 0:03:17.760
<v Speaker 1>others signed an open letter to the u N on

0:03:17.840 --> 0:03:20.960
<v Speaker 1>the subject. In that letter, rather than asking for a

0:03:21.040 --> 0:03:24.000
<v Speaker 1>band the experts applauded the u n's decision to create

0:03:24.040 --> 0:03:27.280
<v Speaker 1>a new Group of Governmental Experts, or g g E

0:03:27.480 --> 0:03:31.240
<v Speaker 1>in the field of lethal autonomous weapons systems. The group

0:03:31.320 --> 0:03:33.840
<v Speaker 1>also extended an invitation to the g g E to

0:03:33.919 --> 0:03:36.280
<v Speaker 1>ask for any sort of technical guidance it might need.

0:03:36.640 --> 0:03:41.200
<v Speaker 1>The letter stated, we entreat the high contracting parties participating

0:03:41.200 --> 0:03:43.720
<v Speaker 1>in the g g E to work hard at finding

0:03:43.760 --> 0:03:46.680
<v Speaker 1>means to prevent an arms race in these weapons, to

0:03:46.720 --> 0:03:50.680
<v Speaker 1>protect civilians from their misuse, and to avoid the destabilizing

0:03:50.720 --> 0:03:54.800
<v Speaker 1>effects of these technologies. The development of AI in general

0:03:54.800 --> 0:03:57.240
<v Speaker 1>has been likened to an arms race. You don't have

0:03:57.280 --> 0:04:00.600
<v Speaker 1>to attach artificial intelligence to a weapon to make it dangerous.

0:04:00.960 --> 0:04:04.000
<v Speaker 1>Plenty of people have expressed concerns about the possibility of

0:04:04.040 --> 0:04:08.560
<v Speaker 1>AI causing harm to humans in general, whether intentionally or otherwise.

0:04:09.000 --> 0:04:12.920
<v Speaker 1>Elon Musk, the entrepreneur behind Tesla Motors and SpaceX, has

0:04:12.960 --> 0:04:15.880
<v Speaker 1>expressed on numerous occasions that we should be cautious with

0:04:15.960 --> 0:04:20.600
<v Speaker 1>artificial intelligence. Meanwhile, Russian President Vladimir Putin has said that

0:04:20.640 --> 0:04:23.719
<v Speaker 1>the nations that advance AI will become the dominant powers

0:04:23.760 --> 0:04:27.080
<v Speaker 1>of the future world. Specifically, he said, whoever becomes the

0:04:27.160 --> 0:04:30.240
<v Speaker 1>leader in this sphere will become the ruler of the world.

0:04:30.960 --> 0:04:33.480
<v Speaker 1>It seems unlikely that the U N will outright ban

0:04:33.640 --> 0:04:36.880
<v Speaker 1>the development of autonomous weaponry. The chair of the first

0:04:36.920 --> 0:04:40.839
<v Speaker 1>Meeting of the Convention on Conventional Weapons is Ambassador I'm

0:04:40.920 --> 0:04:44.160
<v Speaker 1>on Deep Singh gil from India, who has said it

0:04:44.200 --> 0:04:47.599
<v Speaker 1>would be very easy to just legislate a ban. Whatever

0:04:47.640 --> 0:04:50.000
<v Speaker 1>it is, let's just ban it. But I think that we,

0:04:50.200 --> 0:04:53.479
<v Speaker 1>as responsible actors in the international domain, we have to

0:04:53.520 --> 0:04:56.400
<v Speaker 1>be clear what it is that we are legislating on.

0:04:56.960 --> 0:04:58.800
<v Speaker 1>It's more likely that we will see the u N

0:04:58.839 --> 0:05:01.200
<v Speaker 1>try to create a framework in which they will outline

0:05:01.279 --> 0:05:04.719
<v Speaker 1>generally agreed upon boundaries for what is legal and ethical

0:05:04.760 --> 0:05:08.880
<v Speaker 1>in the realm of autonomous weapons. To some, including myself,

0:05:09.160 --> 0:05:11.080
<v Speaker 1>it may seem a little weird to talk about what

0:05:11.279 --> 0:05:15.239
<v Speaker 1>is an ethical versus unethical means to end another human life,

0:05:15.839 --> 0:05:19.760
<v Speaker 1>But that's an entirely different discussion. Maybe someday wars will

0:05:19.800 --> 0:05:22.000
<v Speaker 1>be fought by robots on both sides and the winner

0:05:22.000 --> 0:05:24.279
<v Speaker 1>will be whichever side has the most working units by

0:05:24.320 --> 0:05:26.520
<v Speaker 1>the end of the conflict. But between now and then

0:05:26.960 --> 0:05:30.599
<v Speaker 1>will likely have many discussions about the cold, calculating, and

0:05:30.640 --> 0:05:35.040
<v Speaker 1>potentially terrifying future of warfare. To learn more about artificial

0:05:35.040 --> 0:05:38.320
<v Speaker 1>intelligence and robots, subscribe to the tech Stuff podcast. We

0:05:38.400 --> 0:05:41.039
<v Speaker 1>exp more topics like these in much greater detail. New

0:05:41.040 --> 0:05:44.440
<v Speaker 1>episodes go live every Wednesday and Friday. I'll see you

0:05:44.480 --> 0:05:44.920
<v Speaker 1>against It