Yann LeCun is widely regarded as one of the “godfathers” of AI—but unlike the attire of the famous film gangsters, his bow-tie on Sunday night was an eye-catching shade of red. “I come from a world where corporate attire is a t-shirt and a hoodie, so this is unusual for me,” he told the crowd. “Usually it’s more on the geeky side.”
LeCun was awarded a TIME100 Impact Award on Sunday night in Dubai, in recognition of his achievements in the field over decades leading up to the recent boom in artificial intelligence.
LeCun was awarded a TIME100 Impact Award on Sunday night in Dubai, in recognition of his achievements in the field over decades leading up to the recent boom in artificial intelligence.
Category
🗞
NewsTranscript
00:00 [CROWD NOISE]
00:06 OK, I was scientist.
00:09 I have to tell you, make admission,
00:11 this is the best dressed event I've ever attended.
00:15 [APPLAUSE]
00:20 Usually, it's more on the geeky side.
00:23 I kind of appreciate the recommendation
00:26 of Minister to dress all of us in white.
00:34 I totally subscribe to that.
00:35 I come from a company where the corporate attire is
00:39 a great teacher and a goodie.
00:41 So this is somewhat unusual for me.
00:43 So I've been interested in the mystery of intelligence
00:47 since I was a kid, essentially.
00:51 What is intelligence?
00:52 What is knowledge?
00:53 What is wisdom?
00:54 How do you live here in humans?
00:56 Can we build it in machines?
00:58 Intelligence is the commodity that we are missing the most,
01:01 is in most demand, really, in the world.
01:05 If we were smarter, we would have fewer problems.
01:09 And what we need to do is not just understand intelligence,
01:11 but also amplify it.
01:13 Because amplifying human intelligence
01:16 will perhaps lead to a new renaissance for humanity.
01:20 The way the printing press, perhaps,
01:22 caused a renewal of humanity in the 15th century.
01:30 So as an engineer, I don't think you can understand something
01:33 unless you build it.
01:35 And so trying to get machines to be intelligent
01:38 is a way to understand human intelligence,
01:41 and also to amplify human intelligence.
01:44 And my hypothesis was that intelligence
01:46 emerges from learning.
01:49 Every animal that has a brain can learn.
01:53 I just said that on the video.
01:56 So I worked on learning.
01:59 When I stopped working on learning for a few years
02:02 in the late 1990s, I worked on a technology
02:05 that allowed to disseminate old printed material
02:08 on the internet, because I thought knowledge distribution
02:12 and dissemination was really important for humanity.
02:14 So again, this idea of amplifying human intelligence.
02:18 And I became a professor, because professors basically
02:21 make people smarter, at least if they
02:23 succeed in teaching students.
02:26 So as you heard in the video, in a few years,
02:34 every single one of our interactions
02:37 with the digital world would be mediated by AI systems.
02:41 They will live in our smartphone.
02:42 They will live in our smart glasses.
02:46 And we can ask them questions.
02:48 They will help us in our daily lives.
02:50 Eventually, they'll become smarter than us,
02:52 but they will be still working for us.
02:54 We'll be their boss.
02:56 And we can feel threatened by having
02:59 people who work for us who are smarter than us.
03:01 But as the director of a research lab,
03:05 I only hire people who are smarter than me.
03:08 So I don't think we should feel threatened by that.
03:11 They'll work for us.
03:13 And they will amplify our intelligence
03:14 and the intelligence of all humanity.
03:18 They will constitute the repository
03:20 of all human knowledge.
03:22 And because of that, because our digital diet
03:27 would be mediated by those systems,
03:29 we cannot have a single one or just a few that
03:31 are controlled by a small number.
03:33 It has to be free and diverse for the same reason
03:37 that we need a free and diverse press.
03:40 We can't have a single opinion, right?
03:42 So AI systems will shape our opinion.
03:44 And for that reason, they'll need to be diverse,
03:46 which means the platform on top of which they build
03:50 need to be open source.
03:51 They need to be freely available to everyone
03:53 so that anyone can fine tune those systems
03:57 for their language, their culture,
04:00 their centers of interest, their value system.
04:03 We're going to have a large diversity of AI systems.
04:06 And if we do it right, it's a new renaissance for humanity.
04:10 Thank you very much.
04:11 [APPLAUSE]