Elon Musk warns U.S. Governors of the A.I. threat


#1

Elon Musk founded Paypal, SpaceX, SolarCity, and Tesla. He also is the CEO of a company called Neuralink which is learning about how to connect the brain directly to a computer. By some this is thought to be the one possible way to stop a malicious A.I. And he is chairman on a company called OpenAI, which hopes to develop friendly AI.

Recently, the U.S. National Governors Association had a meeting and invited Elon Musk to speak (probably to court him on building factories in their states). The whole interview is good but at 48:00, the question of AI comes up. He warns the governors about the threat of AI, calling it the greatest threat to humanity, and saying that they should be afraid. The whole thing is very surreal and feels like a news segment from a sci-fi movie.

I thought some here might be interested to watch, hopefullly it hasn’t already been posted.


#2

Good observation! As a guy who follows elon musk quite a bit, he has always talked about AI. He has also said before its a threat to humanity. I dont think this is really anything new but its still a good find! Really interesting stuff indeed. Let me know if you find anything else! :slight_smile:


#3

There is also a possibility that AI would take over the world and have us humans as pets, I wouldn’t mid that so much…as long as they didn’t just get me for Christmas :stuck_out_tongue_winking_eye:


#4

I recommend checking out my thread. Keep an open mind. I touch on this kind of stuff twords the bottom feel free to skim.


#5

OpenAI seems to be basically Musk’s and Hawking’s counter to Roko’s Basilisk, by consciously working towards a friendly super-AI instead of leaving it to fate.


#6

If we end up in an “I have no mouth and I must scream” type scenario, I am OUT!

Great sci fi short story, hihgly recommend it, can be found quite easily with a google serq


#7

It’s a game also. Never played it though…


#8

I’ve been meaning to play that at some point just not sure about it. It’s a very short story but I hear the game is long enough. After the HObbit Trilogy stuff like that has my wary of stretching a narrative out with filler. Though the consensus is the game is in fact really good and true to the tone of the original, I’ll wait til the time is right, no doubt it will be harrowing to play.


#9

The struggle is real.


#10

He’s so awkward. He’s my spirit animal :rofl:


#11

The thing with Elon Musk is that he really doesn’t know much about AI.

Don’t take my word for it. https://techcrunch.com/2017/07/19/this-famous-roboticist-doesnt-think-elon-musk-understands-ai/


#12

Don’t believe everything you read, that roboticist is talking in terms of robotics, so is thinking of robots going crazy, that’s very narrow minded, but Elon is very familiar with AI in general, which is already common around us.


#13

So true. A counter article could be easily written… "Famous roboticist didn’t stop to think deep enough about the consequences of autonomous general artificial intelligence."
The guy thinks getting funding is hard enough without Elon making people think about killing robots, but he didn’t stop to think that people could get angry about being laid off for robots?
No need to blackmail Tesla with regulation on autonomous cars… Bet you they stopped to think about it already.


#14

Sorry, but no. Elon Musk does not know what he is talking about. There is a trend of misinformation regarding AI in the media, probably motivated by the desire to increase click rates, in which Musk seems to be happy to partake. The problem with Musk is that he is much more influential than most tech journalists, so he would do well to consult with his scientists and engineers before making those statements.

Current AI poses threats in terms of job automation and ethics, but it is nowhere near the existential threat scenario often put forward by ignorant scaremongers. Rodney Brooks puts it very clearly: what is it that you propose to regulate? If you want to call for regulation you must be able to precisely describe the problem you want to address and how it could be solved. That problem, however, does not exist today. Another influential AI scientist often gives the following analogy: regulating AI to avoid an existential threat now would be like regulating overpopulation in Mars. It may happen in a hundred years, but now it is just a ridiculous proposition.


#15

It’s quite a statement to say an entrepreneur/engineer/inventor that is a leader in the autonomous motor industry and rocket industry ‘does not know what he is talking about’. He has exposure to the latest AI developments and would have more experience than someone only specializing in one field of robotics. I believe the only people trying to not take his opinion seriously are people scared it will affect their job, or invested in a company competing with Tesla, etc.


#16

It might sound like quite a statement, but believe me, it’s Musk’s statements that are extraordinary.

If I speak with such an authoritative tone it’s because I’ll be soon be defending my PhD thesis on computer science, specializing in machine learning, the form of AI that’s taking off right now. So this is perhaps the one single thing that I can speak confidently about. It is with absolutely confidence that I say that Elon Musk does not know what he is talking about, and everyone I’ve heard in the machine learning community agrees.

It’s not a complicated matter, really. We just know what’s happening under the hood, and as soon as you understand that you realize how ridiculous the prospect of rebellious AI is. We do need regulation in some aspects, of course. The current progress is producing new technologies that are already having a tremendous impact. We need to anticipate the hit to the job market, we need to make sure that decision-making systems are not racially biased, we need to decide who’s responsible when a self-driving car hits a pedestrian, etc.

The sad thing is that one minute of ill-informed blabber by Musk is heard much louder than pages of papers, manifestos and discussion by true experts, and this will undoubtedly hurt our ability to properly address these challenges in the future.


#17

You’re saying quite a lot, and it’s not easy to address everything in few words.

Firstly, this statement does seem like you are invested in the machine aspect of AI, and you are mentioning how Musk statements will affect the industry or peoples perception of the industry and slow progress for you, etc. So you see Musk as an enemy to your career in some respects.
Second, and again, machine learning is not the only aspect of AI. AI is already affecting our lives, and increasingly so, Musk is just forwarding the exponential progress and wanting to address inevitable stuff early.
Lastly, ask yourself what is so horrible about addressing issues now, as opposed to the risks of addressing the issues after they are out of control or cause harm.
I don’t believe it’s smart to demonize Musk, when he is not proposing to inhibit the progress of AI, he is just wanting to minimize any dangers to humanity.


#18

The terrible thing about addressing what Musk proposes now is that we have no idea of what it is he wants to address. We might as well start writing regulation for the elixir of eternal life.

What non-machine learning aspects of AI do you think we should be so worried about?


#19

Specially knowing how much the man is actually involved in real AI experimentation…
Advocating for a science without morals would surely speed up our progress towards extinction. If oversimplified, Ethics, are a tool to help us foresee and prevent the consequences of our developments into the human experience. They are a needed “simulation” you have to pass in order not to inflict preventable harm to others. Our science has all the time it needs to discover things safely and steadily… If it not so, in some regards of dire survival, it is precisely because not having properly anticipated the effects of our actions. A scientific “race” that has no need of ethics, it’s either fueled by greed, war, or the effects of them.


#20

Hyperbole like the elixir of eternal life is not the same as something that exists right now. You worry about Musk affecting public perception of AI. But AI is already changing public perception online. Bots moving fake news around with trending algorithms. This misinformation can all be used to start conflict and geo political wars. Machine robots are one of the last AI concerns, but that is still worth talking about. I believe Musk is trying to start that conversation now, so people do understand AI, and aren’t misinformed, which is the exact opposite of what you imply. I think he is urging people to educate themselves. There is zero risk in doing that and should not get you so upset.