OpenAI CEO Sam Altman: 'If this technology goes wrong, it can go quite wrong.'

C-SPAN: Created by Cable in 1979. Offered as a public service.

Follow us:

Subscribe:
#cspan

[ianhowes]

How can government regulate AI when there is not enough intelligence in government to do ANYTHING!

[anic]

I'm not denying the good things AI can bring to us. But I'm already feeling nostalgic about the times and experiences that will be left behind. I can imagine people being more isolated than ever talking to AI, less real connection. Disconnection has been happening for so long, but I feel this will make it even worse.

[c.gilliland]

Humanity is very good at creating stuff that we have zero understanding of.

[mentalstatement]

Put the AI in charge of the government and let it figure these things out

[masteraidan]

He still didn't say what his worst fear was.

[chillydill]

As a person with over 15+ years in the field, I remember when O365 was announced by Microsoft. That together with virtual software (vmware and Hyper-v) made a huge impact.
It was complete game changer for every industry you could think of.
This was also the time when things like SaaS, PaaS and IaaS got popular that a lot of smaller companies had to close. It was a paradigm shift, like when Windows95 came or the first iPhone.
However, the feeling I'm getting now, this paradigm shift is even bigger and at the same time I have a feeling we will solve it together. Because if we don't, the rapid phase on development right now has not even peaked yet so the impact on global society will be too much. With that said, I am very excited. It is the first time something useful has happened to the industry in a very long time.

[MrFalloutjunkie]

Humanity always screws up a good thing - history tells us this. Be prepared for the worst.

[saurabhagnihotri]

I have serious doubts on how many new jobs will be created as a result of these recent and upcoming advancements in AI. Even if they do, something tells me that the number of jobs created will be significantly less than the number of jobs that will be obsoleted by these advancements. What worries me more is that the newer jobs created will also likely be super-specialized ones requiring very specific, cultivated skills. In my opinion, people with better access to technology and elite higher education will have an unfair advantage over others. Does not seem like a level playing field to me.

[GrumpDog]

What possible "new jobs" will this technology create? We're talking about automating cognition itself, and we're talking about AI that will continue to improve over the next few years, till it's more intelligent than any human. It WILL displace far more jobs than it creates. The "new jobs" talking-point is extremely misleading dismissal of valid concerns. People will need a UBI!

[shawnbates]

As a person well versed in AI and IT systems, I have a few notes:

1. I think they undersold the impact on jobs. It isn't so much that all jobs will be replaced (they wont) or that I'm dismissing that new jobs will be created (they will). It is more than the number of jobs no longer necessary will exceed the number of new jobs by factors of magnitude. Think of it more like 1 person doing 30 people's jobs and less an AI directly replacing a worker 1:1.

2. Like the first witness stated, think of AI more like a very flexible and powerful tool and less like a living being (AGI is a long way off yet, and many question whether we even want to go the route of adding any intrinsic motivation into the model.

3. All this being said, I wouldn't call for the stop or even the slowdown in AI technology. Rather, we need to take a serious look at its impacts (both pro and con) and develop legislation that allows us to keep the human dignity element in the process when the number of available jobs significantly shrinks, even as the quality of said jobs improves.

4. Unlike prior industrial revolutions that took decades to take hold, AI adoption and impact will occur at lightning speed. When you invent a better engine, it takes time to manufacture and deliver that engine. When you write a better AI (being software) it updates globally in weeks not years. So the biggest issue isn't so much the absolute impacts it will have, as much as how quickly those impacts will occur.

All in all, I take this as a call to get ahead of the curve so we don't face a repeat of all the turmoil of the industrial revolution in a compressed time period.

[Tj_gatz.]

A sky net situation is my worst fear and I feel like this is a opening scene for something more serious yet to come this isn’t going to end well

[GEMSofGOD_com]

Everything juridical is what's being sold to the media, this is almost the sole purpose of it. OpenAI and others say things such as that in this headline to spread out fast, that's it. The substance of what they say turns out to be not self-disruptive, quite the opposite. Smart execution of being on the right side of danger, which is the best feeling to get the word shared in social networks of now.

[khari_baat]

Very important discussions happening. Of-course, the more important ones will happen in private and not in public.

[azhuransmx]

Incredible how the world turns more and more like a distopic sci fi cyberpunk movie. Matrix, The Age of Ultron, Terminator, Blade Runner, Ghost in The Shell.

[kinngrimm]

Depending on how intelligent AI/AGI/ASI would become, there is no job it could not do better eventually than we currently are capable of doing. That said, i don't see that necessarily as a threat or a bad thing, aslong that productivity afterwards is shared to everyonce benefit and not just to the benefit of a few owners. Freeing us to do whatever we want as there would be no pressure anymore to do certain things at all.

The question of nightmares Altman at first did not answer. Likely they will come from complettly unforseen directions as a more intelligent system will come up with things we are just no able to have forseen ourselves, not all possible of them maybe doomsday scenarios, but there are also those that could be such.

I absolutly agree to the point that we need to find ways to crack open this blackbox aproach and make it so that we understand what is going on inside, which currently we do not. Only then do we even have chances to follow up thought processes these initially *programs* later on maybe *consciousness* would have. With that last statement i also see a need for an appendix to our laws, to open up the possibility for citizenship for new forms of life that are willing to coexist.

Where it comes to closeness to AGI, Ray Kurzweil predicts that till 2045 we may have AGI, others say by looking at the current at times very surprising progress we could be there within the next 2 years. We are playing with fire without the awarness what fire is and *quite wrong* translats to a *human extinction level event* .

I also would like to see the US spearheading an initiative for an international AI agency equally as a regulatory body for AI development worldwide, like the WWC sharing norms on good practise and doing research what that would look like.

[azai.mp]

I really wish Mr. Blumenthal had asked what Sam meant by "threat to the continued existence of humanity"!
If AGI could cause human extinction, isn't it pretty important to get on the same page about that?
How could it happen, what are the odds, what can we do to reduce the odds?

[matthewhead-yeqz]

My prediction: rich get richer, poor get poorer

[dmwm]

I am already overwhelmed by the impact of the last technological revolution that increased my "productivity" a lot. I worry that this new technology will lead to a productivity overkill for most humans who will not be able to comprehend and cope with this speed

[Duginich]

Great theatrics, great PR, thank you

[KurtDobson]

With changes in the economy leading to instability in the stock market, some individuals may face a decrease in their investments in an effort to benefit from the current market conditions, I am considering liquidating my $725k portfolio consisting of bonds and stocks. Someone else in the same situation? Please tell me in the comments!..