WEBVTT - The AI Arms Race

0:00:03.720 --> 0:00:07.920
<v Speaker 1>Some say artificial intelligence will transform our planet. Vladimir Putant

0:00:07.960 --> 0:00:11.440
<v Speaker 1>says harnessing AI is the path to ruling it. I'm

0:00:11.520 --> 0:00:18.480
<v Speaker 1>John and Strickland, and this is tech stuff daily. Artificial

0:00:18.480 --> 0:00:21.360
<v Speaker 1>intelligence is a hot topic of discussion. While it might

0:00:21.400 --> 0:00:25.280
<v Speaker 1>conjure up thoughts of humanlike robots or super intelligent machines,

0:00:25.560 --> 0:00:29.120
<v Speaker 1>the term artificial intelligence is actually a really broad one.

0:00:29.480 --> 0:00:33.080
<v Speaker 1>It encompasses everything from basic functions of intelligence such as

0:00:33.159 --> 0:00:36.000
<v Speaker 1>voice recognition, and goes all the way up to concepts

0:00:36.080 --> 0:00:40.520
<v Speaker 1>like cognition and self awareness. Perhaps because of this broad definition,

0:00:40.720 --> 0:00:44.360
<v Speaker 1>discussions about AI are also all over the map. Some

0:00:44.440 --> 0:00:47.000
<v Speaker 1>people tout the ability of AI to help us accomplish

0:00:47.040 --> 0:00:50.479
<v Speaker 1>difficult tasks or to offload more tedious work onto machines.

0:00:50.840 --> 0:00:54.120
<v Speaker 1>Others point at AI as being a potential existential threat

0:00:54.160 --> 0:00:58.280
<v Speaker 1>for all of humanity. Elon Musk, the entrepreneur behind such

0:00:58.320 --> 0:01:01.840
<v Speaker 1>companies as Tesla and space X, has famously spoken out

0:01:01.880 --> 0:01:06.120
<v Speaker 1>against developing AI without some sort of strong guiding principles.

0:01:06.440 --> 0:01:09.840
<v Speaker 1>Musk's argument is that AI poses an enormous risk to humans.

0:01:09.880 --> 0:01:12.120
<v Speaker 1>It doesn't have to be super intelligent to be a

0:01:12.160 --> 0:01:16.640
<v Speaker 1>problem either. Take the concept of weaponized AI, for instance,

0:01:16.959 --> 0:01:20.200
<v Speaker 1>This could take many forms. One example might be military

0:01:20.280 --> 0:01:23.920
<v Speaker 1>robots with fairly simple AI. That AI could look for

0:01:24.120 --> 0:01:27.360
<v Speaker 1>enemy soldiers or vehicles and direct a robot to fire

0:01:27.440 --> 0:01:30.360
<v Speaker 1>upon them. Such a weapon would be a powerful tool

0:01:30.400 --> 0:01:32.720
<v Speaker 1>for a nation to use against enemy forces, and it

0:01:32.800 --> 0:01:36.600
<v Speaker 1>might mean that countries possessing that capability would be more

0:01:36.640 --> 0:01:40.200
<v Speaker 1>willing to enter into conflicts. After all, the soldiers are

0:01:40.200 --> 0:01:43.679
<v Speaker 1>made out of metal, glass, and plastic. They're expensive, but

0:01:43.720 --> 0:01:46.920
<v Speaker 1>they're disposable. If you can put robots against an enemy

0:01:46.959 --> 0:01:49.200
<v Speaker 1>and keep your people safe, you might be more willing

0:01:49.240 --> 0:01:52.840
<v Speaker 1>to enter into that armed conflict. Or you could develop

0:01:52.920 --> 0:01:57.720
<v Speaker 1>artificial intelligence designed to infiltrate and sabotage enemy infrastructure. Rather

0:01:57.800 --> 0:02:00.760
<v Speaker 1>than controlling a physical robot, these program ms would attempt

0:02:00.760 --> 0:02:04.800
<v Speaker 1>to penetrate, spy upon, or disrupt systems like power grids

0:02:04.880 --> 0:02:09.040
<v Speaker 1>or other critical infrastructure. Security firms like Semantech have already

0:02:09.040 --> 0:02:12.959
<v Speaker 1>warned the public that hackers had successfully compromised several energy

0:02:13.000 --> 0:02:16.920
<v Speaker 1>companies to an alarming degree, and that if hackers so wished,

0:02:17.000 --> 0:02:20.600
<v Speaker 1>they could potentially cause massive blackouts across major sections of

0:02:20.600 --> 0:02:23.960
<v Speaker 1>the United States power grid. And automated intelligence that can

0:02:24.000 --> 0:02:27.480
<v Speaker 1>behave in a similar way would be incredibly dangerous. None

0:02:27.520 --> 0:02:30.560
<v Speaker 1>of this requires the artificial intelligence to have anything remotely

0:02:30.639 --> 0:02:34.080
<v Speaker 1>similar to human intelligence, but if it is possible to

0:02:34.120 --> 0:02:37.600
<v Speaker 1>make machines think for themselves, that could be even more risky.

0:02:37.880 --> 0:02:41.280
<v Speaker 1>Russian President Vladimir Putin said that whoever becomes the leader

0:02:41.360 --> 0:02:45.520
<v Speaker 1>in this sphere, that sphere being artificial intelligence, will become

0:02:45.560 --> 0:02:48.760
<v Speaker 1>the ruler of the world. Puton seems to suggest that

0:02:48.840 --> 0:02:51.560
<v Speaker 1>AI will be the next big arms race, and there

0:02:51.600 --> 0:02:54.200
<v Speaker 1>are a lot of contenders pushing toward developing ever more

0:02:54.280 --> 0:02:58.280
<v Speaker 1>powerful AI besides Russia, including China and the United States.

0:02:58.840 --> 0:03:01.799
<v Speaker 1>Over at the United Nations, discussions about weaponized a I

0:03:01.960 --> 0:03:05.359
<v Speaker 1>have begun. A group of more than one technology experts

0:03:05.400 --> 0:03:07.400
<v Speaker 1>send a message to the U n to call for

0:03:07.480 --> 0:03:10.800
<v Speaker 1>a ban on autonomous weapons. The message, in part read

0:03:10.840 --> 0:03:13.360
<v Speaker 1>that if development in that space continued, it would be

0:03:13.480 --> 0:03:16.880
<v Speaker 1>very difficult to reverse direction later on. In the meantime,

0:03:17.080 --> 0:03:20.840
<v Speaker 1>weaponized AI could pose an enormous threat to vulnerable populations,

0:03:20.880 --> 0:03:23.880
<v Speaker 1>and that's assuming everything is working properly in the first place.

0:03:24.200 --> 0:03:27.359
<v Speaker 1>The United Nations has a rescheduled meeting of government experts

0:03:27.400 --> 0:03:31.280
<v Speaker 1>to talk about AI that will happen in November. The

0:03:31.320 --> 0:03:33.960
<v Speaker 1>meeting was rescheduled because some of the member nations had

0:03:34.040 --> 0:03:36.640
<v Speaker 1>yet to pay their dues in September when the meeting

0:03:36.680 --> 0:03:40.080
<v Speaker 1>was originally supposed to take place. While AI could potentially

0:03:40.080 --> 0:03:42.800
<v Speaker 1>cause a great deal of harm, there's no denying it

0:03:42.800 --> 0:03:45.520
<v Speaker 1>could also have a positive impact on the world if

0:03:45.720 --> 0:03:49.920
<v Speaker 1>used responsibly. For example, IBM S Watson platform has been

0:03:49.960 --> 0:03:53.680
<v Speaker 1>working with healthcare professionals to help doctors diagnosed and treat patients.

0:03:54.040 --> 0:03:57.440
<v Speaker 1>Autonomous vehicles have the potential to prevent thousands of deaths

0:03:57.480 --> 0:04:00.880
<v Speaker 1>from car accidents caused by human error. Abots are taking

0:04:00.920 --> 0:04:05.360
<v Speaker 1>on more jobs that traditionally are dangerous for humans to perform.

0:04:05.400 --> 0:04:08.520
<v Speaker 1>There are many applications of AI that are clearly beneficial

0:04:08.520 --> 0:04:11.400
<v Speaker 1>to humans as a whole. The question about whether AI

0:04:11.480 --> 0:04:14.200
<v Speaker 1>will ultimately be a boon or a threat to mankind

0:04:14.280 --> 0:04:17.120
<v Speaker 1>seems to hinge largely upon how we are determined to

0:04:17.200 --> 0:04:20.839
<v Speaker 1>implement it. If, as the group of technology experts argue,

0:04:21.040 --> 0:04:24.159
<v Speaker 1>we place strict limitations on the use of AI in

0:04:24.240 --> 0:04:28.440
<v Speaker 1>military applications, we may say that artificial intelligence is the

0:04:28.480 --> 0:04:31.719
<v Speaker 1>most important technological advanced human beings have made since the

0:04:31.720 --> 0:04:35.480
<v Speaker 1>invention of fire. If we do not heed those warnings,

0:04:35.880 --> 0:04:38.719
<v Speaker 1>we may say it's the most destructive tool ever created.

0:04:39.279 --> 0:04:42.080
<v Speaker 1>To learn more about artificial intelligence in all its forms

0:04:42.120 --> 0:04:45.600
<v Speaker 1>and applications, subscribe to the tech Stuff podcast. The show

0:04:45.640 --> 0:04:48.320
<v Speaker 1>publishes every Wednesday and Friday with a deep dive on

0:04:48.400 --> 0:04:51.840
<v Speaker 1>big topics in the technology space. Also, I throw in

0:04:51.880 --> 0:04:55.080
<v Speaker 1>a lot more jokes. That's all for today. I'll see

0:04:55.080 --> 0:05:03.880
<v Speaker 1>you again soon. Wal Wal won