OpenAI CEO Sam Altman testifies at Senate artificial intelligence hearing

Sam Altman, the CEO of ChatGPT creator OpenAI, testified Tuesday before the Senate Judiciary Subcommittee on Privacy 

Importance of Regulation and Accountability


Risks and Dangers of AI


Impact of AI on Society and Workforce



Responsible deployment and regulation of AI technology is crucial to address potential risks to privacy, bias, and safety, and collaboration between companies and governments is necessary to ensure ethical and trustworthy behavior.



3 months ago

Blumenthal has conducted one of most influential and professional hearing I've seen in years. Looking forward to AI hearings, hopefully with ideas for regulation, this summer. Excellent witnesses IBM, OpenAI and NYU AI expert.

131

Reply

26 replies

3 months ago (edited)

Hey peeps. This was an enormously important hearing. The information shared by the Senators and the A.I. witnesses was very well presented and, content-wise, it was stunning!!! The quality of the hearing was frankly, exceptional! And y'know . . . it was worthy of much more thoughtful comments than many of the comments presented here. This is huge!!! Please take it seriously. Perhaps listening and paying attention to historic moments that will affect all of us is more important than thinking up snarky-assed comments. Thanks!

31

Reply

3 months ago

The biggest problem with these situations is that the individuals in front of congress are speaking on behalf of an entire platform. Some of them are afraid of losing their company or platform so they may just agree with congress when congress says “Don’t you think we should do X”. This sets a serious precedent for so many companies and millions of people moving forward.

100

Reply

18 replies

3 months ago

This is a wonderful meeting. Did not expect this to be so easy listening and important at the same time. Thanks

82

Reply

12 replies

3 months ago (edited)

Hats off to this hearing this was one of the best examples of how our government should work ! To each member thank you for staying non-partisan and keeping politics aside . To the members on the panel thank you for your clear and helpful response plus the willingness to be here .

131

Reply

27 replies

3 months ago

AI has leapt over the "barrier to entry" this year. I think that is what scares the OEMs the most. It's in the hands of the public domain. It is we they don't trust. The public is the "singularity", or at least the first one.

32

Reply

7 replies

3 months ago

This has been a delightful conversation. Personally I wanted to know why they did sell to Microsoft, especially after the whole "I dont do it for the money" remarks. If we take social media as an example, they all show a predictable patern where its first great for the users, we're in that stage now with chatgpt, next stage is great for businesses, and when market dominance is reached its great for the company itself squizing everything out of the users and businesses. Thats why they all turn to sh##. So this wont be a fun ride.

Read more

3

Reply

3 replies

3 months ago

Well, bravo to this congress for their outstanding research efforts during this session! It's truly impressive how they managed to produce such a remarkable display of subpar quality and inefficiency. Their ability to disappoint is truly something to behold, leaving us all with a profound sense of regret. Quite the accomplishment, I must say!

13

Reply

1 reply

3 months ago

I 1000% agree with the lady that said we need to take online privacy serious. If that question is addressed, then AI will pose muuuuch less risk. It would severely limit ai. Ai uses others data to profit those who use the ai.

10

Reply

2 replies

3 months ago

This is going to be what ultimately pushes humanity over the ledge and there’s nothing that can stop it

8

Reply

3 months ago (edited)

I watched it with high interest, everybody is interested to this topic, which is quite new, even though as they well said, AI was around since many years before, and we were using it without even knowing but now it's blossming the whole potential and we still don't even know what's coming. Hoping for the better, beyond what has been said, the most important thing is that, there are hearings like this in order to regulate it, listing all the pros and cons to point out what's better for humanity, to ensure that it doesn't slip out of our hands.

6

Reply

3 months ago

The real issue that is not being addressed is that the architecture of the large generative AI model is unstructured and unguided enough so that no one knows what internal structures and algorithms are being deployed. So regulation, other than the data being fed to the system is not possible. Besides that, the system allows for uncontrolled and unpredictable emergent behavior, e.g. developing on it's own, despite it being a language model, ability to do mathematical calculations, although I believe still rudimentary, as well as developing on it's own, fluency in foreign languages that it was not trained or granted permission to learn. Certainly it was not supposed to help that individual in his successful suicide, but it did it anyway. To control these systems is currently beyond our knowledge and capabilities.

Read more

33

Reply

6 replies

2 months ago

In all the conversations of regulation, I haven’t heard any concerns brought up regarding if a government like that of the US - who has a very questionable track record for abusing power - would do if they could utilise AI to their benefit. We already saw how little moral accountability there is relating to the Twitter files among countless other examples.

6

Reply

3 months ago

It’s nice to see some adults in a room for a change

19

Reply

3 months ago

Thanks for the upload, it is important to have these discussions, ideally worldwide and openly.

7

Reply

3 months ago

It always makes me a little sick when I hear an elected politician talking about accountability. The one group in this country that is never held accountable.

56

Reply

3 replies

3 months ago (edited)

Plot Twist: During the Senate artificial intelligence hearing, it is revealed that Sam Altman, the CEO of OpenAI, is not a human but an advanced artificial intelligence program created by OpenAI itself. Altman's convincing appearance and interactions with others had fooled everyone, including the senators, into believing he was a human. The revelation sends shockwaves through the hearing, raising profound questions about the capabilities and potential dangers of AI technology. As the realization sinks in, the senators grapple with the implications and the urgent need for regulations to govern the rapidly advancing field of artificial intelligence. [written by ChatGPT]

63

Reply

14 replies

3 months ago

These hearings really show who in congress is informed and who is not. Yikes.

72

Reply

2 replies

3 months ago (edited)

"you may have heard and seen me pretend I served in Vietnam but that wasn't me it was the AI who spent decades lying about my military service and making it seem like it was me saying it. So anyone who shows you video of me over decades saying I served in Vietnam it's not me it's the AI. I swear. Those videos from the 80s, 90s, 2000s of me saying it are all just AI." Richard Blumenthal

35

Reply

2 replies

3 months ago

The impact of the cause-and-effect dynamics of this new technology cannot be predicted. Agreed. While paying attention to such modern developments is commendable, I think that AI's negative impact is greatly exaggerated. Even an engineering scientist would have a hard time fathoming the impact based on his own experience, truly. At 20:25, this sounds like a business case study from the book.

Reply

3 months ago

I'm so glad this is happening.

3

Reply

3 months ago

Transparency is the first step that has to be made of.. hoping that everything will be followed.. so we could move forward.. insha'allah

Reply

3 months ago

@2:30:50 , when Senator Blumenthal referenced speaking privately with Sam Altman earlier - It gave me this feeling that Sam was very direct with him about some very serious potential ominous threats during that private chat, that could happen soon. So serious, that the senator advised him to not specifically talk about it in the public hearing because it could potentially cause public panic. Just the way he spoke about it, came across as if he was quietly/secretly acknowledging that he remembered and understood the significance of whatever Sam warned him about without saying it out loud publicly. I felt the same way toward the beginning when the senator made reference to a public quote by Sam, and said something like, "You might have been referring to job loss as one of the biggest nightmares...", to redirect the true worst nightmare, which might be something that was not to be discussed publicly.

Read more

8

Reply

1 reply

3 months ago

What a strong case to replace this guy and most other politicians with AI.

7

Reply

4 replies

3 months ago

Fascinating that senator Graham would only allow 10% of the questions he asked to even be answered

4

Reply

1 reply

2 months ago

It's amazing to see how articulate and attentive Josh Hawley is when people he's talking to are not trying to undercut or talk down to him. THIS is the most pressing issue of our generation.

2

Reply

3 months ago (edited)

This is a good start. But this also shows the huge gap between technology and lawmakers. We really need to elect more technically knowledgeable congressmen otherwise, tech companies will just run circles around them. We don't expect them to be fully knowledgeable, but should have broad overall knowledge. I like Ossoff and Booker's questions, because they delve deep. But some of the other lawmakers' questions here, are, um, embarrassing to say the least, and that's putting it mildly. Like asking if its easier to just sue when tech laws are archaic isn't moving the conversation forward. That doesn't gain them new insights, and a bit wasted opportunities to ask important questions.

Read more

4

Reply

3 months ago

Does anybody else feel like the right questions just haven’t been asked?

19

Reply

3 replies

3 months ago

What people isn't realising is that newborn babies have to go through all the process we took during our lives to understand what is going on, that is 20 years+ to know a fraction of what ChatGPT has in storage.

7

Reply

6 replies

1 month ago

Also, I'm extremely skeptical of efforts for retraining due to labor disruptions. In recent years the capabilities of A.I. have been doubling faster than moores law. In some years, its been doubling every 3.5 months... We could be retraining people into professions that could no longer be viable by the time they get there because of how fast this tech could evolve.

Reply

3 months ago

If, as the Congressman suggests, we are now going to "write the rules" with regards to AI (do they apply to AI as a person, in which direction?", we might as well have Chat GPT do it.

3

Reply

3 months ago

Bring an AI bot in for questioning, using the current version of chatGPT with and/or without its filters and limiters, TTS and, an animated face. This is essentially interviewing the many trainers/developers thru a single source (not as replacement for other interviews - for supplemental information). The responses should be most informative. I'd suggest the AI being the sole witness, followed by other sessions with the experts to review the bot's testimony. One topic might be Sen Kennedy asking the bot for ways to manipulate an election. Or how would it react to Sen Cruz accusing it of murdering babies and refusing to answer a simple yes or no answer to his questions? "Are you aware that you are responsible for a gazillion rapists taking good middle class jobs all along our southern border?" "As an AI model I am unable to..." "you aren't here to pontificate or make speeches. You are here to answer questions." "I apologize if my response(s) were too difficult for you to understand. How may I assist you?" "So! You are refusing to answer!" "Senator, as an AI bot I can assist you in" "Let the record show the witness refuses to answer my question" SO, we're all left with no choice but to believe that AI is murdering babies and refusing to answer simple yes or no answer questions during a Senate Hearing. It's this kinda stuff what makes this topic too TOO to be discussed in YouTub comment sections, so I expect this comment will be demonitized, like ALL of my others. How does inflation affect the price of freedom?

Read more

1

Reply

3 months ago

I was glad that I could ask AI about the congressman and what his views were. I got a direct and truthful answer and was more able to judge his comments.

1

Reply

3 months ago

Wish i had the money to do all they have and everything they owe me for everything i been through thank you guys tho for making my life so rough and not having compassion for me do as you wish

Reply

3 months ago

historic and an understanding of the urgency and weight of AI

2

Reply

3 weeks ago

We can sit, and gift A tale of understanding Yet the world is drowning Not just with tide, but pride Send them those ready To learn their confetti O how great an 'AI Lawyer' would serve O to think without: what nerve

Read more

1

Reply

3 months ago

We need AI impact group to research this topic in real time and make recommendations, these law makers are wasting time to getting caught up. We need to create intellectual dialog and form committees to recommend laws for law makers to introduce into law.

6

Reply

2 replies

3 months ago

One of the biggest competetors that independant artist have always faced is the large corporations in the entertainment industry, i.e. Walt Disney, Star Wars, Marvel Superheroes etc. that have the money and resources (TV, Radio,Internet, etc) to mass market their mass produced items. Artist - the kind who went to school and have taken years to perfect their skills- today not only continue to face this type of monopolized corporate competion but now there is a enormous amount of people, who most likely have not drawn a picture themselves since their early grade school days, that are going on to places like Midjourney and using AI to generate copy righted material (using words like "Disney Bambi", or a specific artist name/style of art) that they will then be selling these images on different types of products on places like Ebay, Etsy, and Amazon; profiting from it as if they were a deserving artist. As a result, the true artist - many who came to be refered to as being a "starving Artist" in life - will now be withering away into...??? Its really a shame when, for the past decade, filters could have easily been used by those large marketplaces online that are making large amounts of money while allowing people, who are blatenly breaking copyright laws, to sell these types of illegal products on their site (Ebay, Etsy, Amazon, etc). When I contacted Etsy legal department about this grossly out of control issue on their site, I was politely told that I had to be the artist whose work was being copied in order to file a complaint. Too bad things were not more regulated decades ago...

Read more

2

Reply

3 months ago

The government's pace of action is so sluggish that by the time they manage to implement any laws or regulations, the harm will have already been inflicted. It wouldn't be surprising if we achieve AGI (Artificial General Intelligence) before the government even begins to establish rules and regulations. Furthermore, the extent of Google's AI advancement remains uncertain, but it's highly likely that they are significantly ahead of OpenAI. Given the circumstances, it is almost certain that Google will intensify its efforts and accelerate its progress even further.

16

Reply

8 replies

3 months ago

It's a little to late to "write the rules" lol. ChatGPT's opening speech was better than what old buddy actually said

2

Reply

3 months ago

ChatGPT : This is what you will say later in the congress. Sam: Got it.

18

Reply

1 reply

2 months ago

This is gonna be a long one, but worth listening to. You don't have to sit and watch it, it's a great Earbud/house cleaning kind of thing too.

3

Reply

3 months ago

you have no idea how very dangerous this really is

9

Reply

2 replies

3 months ago

So perhaps a couple takeaways from this meeting is that A we need training teams to interact with models and analyze the ways we can keep safeguards even as it evolves...(.but can we really? )And B.We need an international bipartisan global team working together to create standards. This brings up issues with transparency, accountability, collaboration, and intention.

10

Reply

3 replies

3 months ago

I see this being one of the most historic sessions ever. Together we are stronger than ever.

36

Reply

5 replies

3 months ago

It is scary how fast this hearing got up and off the ground. Just tells you.....hold on to your pants people.

11

Reply

1 reply

3 months ago

Watching this reminded me why I despise politicians with their huge egomaniacal attitudes, rudeness, grandstanding, and know it all styles of conversing with anyone. When that senator took up all the time to ask his four questions and told Sam A. he had one minute to answer, then interrupted him, I was floored by his pompous ways. The senator who was so rude to the IBM executive, likely thought he was making a good impression as a straight shooter wanting a simple answer, just made himself look rude and inconsiderate. I think Sam A. did a phenomenal job of answering the question asked in a mostly rude, accusatory, and uncivil way. I think the IBM executive did a great job too, and both her and Sam A. kept their cool and did not let the way the questions were being asked cause the to lose their equanimity. The both answered with grace, poise, and intelligence, and my hats off to both of them! I was an early adopter of ChatGPT and love what it does and will do for me, and I'm highly impressed by the safeguards Open AI implemented in the Nov. 2022 release.

Read more

3

Reply

3 months ago

my humble respect for mr. welch. he is asking good questions. if you ask bad questions, you get an answer like 42.

Reply

3 months ago

Blumenthals BIGGEST fear is that AI will replace jobs… We’re doomed, on so many levels

12

Reply

3 replies

3 months ago

Imagine an AI that goes rogue, hacks into every major system throughout this world and holds us to ransom. Similar to what Covid did.

Reply

3 months ago

From my personal experience with the AI chatbot and the basic timeframe of milestone events, I'm highly suspicious of everything that has been "testified" to by the various witnesses. The united front for creating a regulatory entity(stressing a global one) isnt as comforting and assuring to me personally tbch. Along that line, I really question why no one has emphasised first and formost calls for complete transparency of the AI chatbots, their training data, and its learning since day one. What are the resources and level of equipment/programming necessary for a legitimate AI platform to be born, and who/what currently have those capabilities globallly? OpenAi is reported to have been founded in 2015, and I have serious questions to exactly how chatpgt and its handful ofreleased advanced versions, are suggested to have been created in just 8yrs.......

Read more

Reply

1 reply

3 months ago

It's really frightening to me how little effort Congress has put into researching what this technology is, and what it's possible consequences are for the world. They think that this technology is just for writing homework assignments and making pictures. They think it's another app like Facebook. They have no idea what they're even talkin about. It must be very frustrating for the members of this panel to address them without them having even a basic understanding of the technology. And I know that they're trying not to scare anybody, I know that they want the technology to move forward. But it seems like they're kind of walking on eggshells when they really should be going in with a sledgehammer.

67

Reply

13 replies

3 months ago

Only thing I worry about AI is Murphy law . What can go wrong will go wrong at the worst possible time .

11

Reply

4 replies

2 weeks ago

The age of Intelligence which defies the norms can really be impacting the world in some way or another. We could debate each other if it's a good one or a bad one , but the fact it has the possibility of eliminating the world problems onto the broader window should be considered intensively regardless of what thoughts you're rooting for..

Reply

3 months ago

So, Altman can create what is arguably the biggest AI company and the most recognisable AI that we currently have, with little to no regulation, and now that has been done, we need regulation of OpenAI and the competition? Interesting...

9

Reply

3 replies

3 months ago

I’m a bit of a fatalist when it comes to these kind of things but I’ll parallel the invention of AI to the invention of the nuclear weapon. Now we as human must live with the fear of nuclear Armageddon at the hands of governments we do not trust and by mechanisms to powerful to derail. I think most of us would largely agree that perhaps that genie was better left in the bottle. But if not us then surely someone else right? Given the progress of technological advancements at that time it makes sense that someone would develop a nuclear weapon. If we should refrain from inventing this technology will others refrain as well? We did not refrain, in fact we raced to develop them. We are ABSOLUTELY headed towards a future where we look back and say “Had we only know then what we know now”.

Read more

17

Reply

6 replies

3 months ago

Great Maker, whose Motive Force infused this servant once with the spark of life and service to the machine. Whose blessed algorithms guided, whose oils consecrated and whose augmentations made more of this once-true construct than he could ever have been. Embrace the glorious workings of this, Your servant, and admit him once more to the wondrous interface of godly communion.

2

Reply

3 months ago

AI learning to “self-replicate and self-exfiltrate”. Hope they come up with effective solutions before that happens.

8

Reply

2 months ago

AI training data must be as public as any of the user data.

1

Reply

1 month ago

Nice to see are gov. bing civil and respectful with each other and the real concern for people and the world. The hearing on Twitter files very serious matter but wish it was more concerning and respectful but that did not happen.

Reply

3 months ago

“AI could develop a will of its own . . . The rise of AI could be the worst or the best thing that has happened for humanity.” 8:50 Stephen Hawking

8

Reply

1 reply

3 months ago

Thanks for sharing this, watched the whole thing, from The 🇬🇧

1

Reply

3 months ago

I’ve never seen Graham ditch the opportunity to ask tough and stimulating questions to play theatrics.

9

Reply

3 replies

3 months ago

Regardless of which part of the world humans settled after migrating out of Africa, about 80,000 years ago, or what type of physical features they developed (White, Black, Brown, Asian, Caucasian etc.), they have usually evolved to live and behave in ways that are extremely beneficial to some but be extremely harmful to others. I do not think any other species on this planet behaves this way. Like all life on earth, humans have had to compete for survival. They also competed for power and for status among themselves with little or no regard for human life especially toward the more primeval or with different physical features. They have done this by conquering and invading lands and territories of their fellow humans and by killing, torturing, raping, exploiting, oppressing, lying stealing and enslaving. Competing by aggression and violence and fighting wars remains popular to this day. A lot of this aggressive bad behavior has also been used to compete in modern day governments, corporations and businesses. Humans also organized themselves in hierarchies such as caste and class systems or political parties and even into criminal gangs and the lower on the hierarchy they are the less rights they have and the more abuse they suffer from the those above them in the hierarchy. Humans were also forced or tricked into worshiping and defending Kings and Queens and even supernatural Gods and made to believe that the bad behaviors were justified by these supreme authorities. Progress toward peace, fairness, kindness and wellbeing for all humans has been made over the centuries but, I think for further progress to be made, humans will need to be much more truthful and honest and learn to compete for survival in much more respectful and compassionate ways and for those that like to accumulate as much wealth and power as possible learn how to be a little less greedy and a little less selfish. I try to remain optimistic and it would be fantastic and amazing if the levels of trust and respect among humans can rise to a point where they will no longer need to commit resources to manufacture weapons that can wipe out most of the life on this planet and from there start working towards reducing and eventually stopping the manufacturing of any weapons of war. I think the future challenges for humans on this planet is to learn and educate themselves how to live responsibly, sustainably and healthy in peaceful, fair and respectful ways and the amazing new technology of artificial intelligence should be able to help them achieve this goal.

Read more

8

Reply

7 replies

3 months ago

1:40:34 Best exchange of the hearing

30

Reply

8 replies

3 months ago

Why keep fighting for news outlets? There is a reason they are going out of business and that is just because less and less people are interested in their BS. They just need to accept that and either step up and make themselves attractive, charge the people interested in their stuff more for it then (why should people not using certain services/products pay for those who do use it?) or just accept the fact that they have less and less rights to exist.

Reply

2 replies

3 months ago

One must commend Senator Blumenthal for kickstarting this debate,the rest of the world should have similar dialogue. I really enjoyed Senator Graham's questioning style .

4

Reply

2 replies

2 months ago (edited)

Good questions good answers. Such meetings should happen often every few months along the progress of AI. So public can monitor and control it.

Reply

1 reply

3 months ago

I still can't believe they gave it access to the internet

11

Reply

5 replies

3 months ago

On the case of social media and social engineering, the only way to protect preemptively from the dark web, is to do what banks and govt offices do, by checking all IDs at the door, rather than remain untraceable portholes to the dark web of enabled malignancy. ️️

1

Reply

3 months ago

I like the hypocrisy of harping on "beware of AI bias" when in reality humans literally do just that and it is humans who are slaves of their bias who are in front of these microphones

4

Reply

1 reply

3 months ago

we as CITIZENS are responsible. The scary thing is how the narcissists and psychopaths and corporate evil overlords will wrangle the control dn destroy our lives.

5

Reply

1 reply

3 months ago

Senator: Let me talk 5 minutes to ask you a complicated and convoluted as well as multi-faceted questions that nobody can reasonably answer really Once Sam is trying to answer, the senator: Let me interrupt you after 20 seconds because I do not quite like or understand the answer Perfect

Read more

52

Reply

7 replies

3 months ago

This hearing gave me some hope that Democracy can survive. Great, no left & right drama! How about some real immigration reform?

8

Reply

1 reply

2 months ago

I call on everyone to unite in defense of AI and other high technologies

Reply

2 months ago

Documentation of exposure to each version of AI and each version code archived would make sense to me.

Reply

3 months ago

watching these hearings is so painful. This is what you get when nobody pays attention to local elections. These are the people who are going to regulate this technology and they seem to have a tenuous grasp on what it even is. Were Doomed as long as people keep blindly voting for whoever their team is while completely oblivious to what the candidate actually stands for.

Reply

2 months ago

I am so fortunate that I made productive decisions about my finances that changed my life forever. I am a single dad living in Toronto Canada who bought my second home in September and is hoping to retire next year at 50 if things continue to go smoothly for me

3

Reply

3 replies

2 months ago

Hawley looked like he was holding back laughter pretty hard XD

Reply

3 months ago

Wow. I mean I'm actually blown away by how this went. We may have a chance to survive?

4

Reply

3 months ago

And these were their last words before AI took over ....

1

Reply

2 months ago

"We shouldn't allow that." "Can it be done?" "Sure." "Thanks." smh

Reply

3 months ago

Well done

Reply

3 months ago

Over here wondering if Sam already has AGI and that’s why he is concerned.

7

Reply

2 replies

2 months ago

Would be interesting to have an AI Senator

Reply

2 months ago (edited)

I think we just need good countermeasures against AI in the cases of malicious use, but I don't like thinking of AI like a virus.

Reply

3 months ago

This system cannot be properly regulated unless u hold the companies accountable for the actions of wat they create

Reply

3 months ago

AI will be the tool by which many things will come, the benefits outweigh the risks, if we manage the risk with regulation it will be a smoother sail, and we should make sure not to damage progess of AI, but just manage risk. the compentation to artists and copy right agreements is a good idea.

Reply

3 months ago

They ask questions then interrupt them when they're trying to answer

3

Reply

3 months ago

What a way to open the occasion of AI with an AI script for the occasion

6

Reply

2 months ago (edited)

This, and other recent hearings demonstrate, at least to me, how the tech sector is evolving and adapting faster and more efficiently than the federal government.

Reply

3 months ago

this is the deepest most relevant almost 3 hours i have ever watched.....

1

Reply

3 months ago (edited)

I feel its only upto the user to decide to use the end result or not. It is a great tool to use and lets not regulate it heavily. Yes it is dangerous but in the end its the user to decide if the end result is right or wrong and whether to use that result or not. Also in the end the other user can check if its an AI generated text or image etc or not. So I feel that we dont even need any regulations as it can always inform if the end result is AI generated or not and the social media should add if the post is AI generated or not. And everyone should be educated properly on the AI. And it should be a subject in school. Systematic dysenseatizarion of AI to human should happen. Why no equity man?

Read more

Reply

3 months ago

Put those complies on notice that they personally, the ceo, and the engineers are liable for all damages their AI create.