Yeah.

Uh, I think the vast majority of people who are working in the AI space actually want there to be regulations.

They're the people who are closest to the technology.

They see how powerful it is.

And so both in terms of contributions to me, but in contributions to groups like Guardrails Alliance that exist to counteract Leading the Future, you're seeing a lot of people at all of the labs say, "Yeah, we, we want regulation." This is a intellectually challenging problem that we're building something really powerful, but the question of who decides how it should be used and what the limits are shouldn't be made in s- just in Silicon Valley or just by their, uh, CEOs.

Now you have some of the executives who don't want there to be any regulation whatsoever, and so they're willing to spend big in order to get that.

You also have Chris Lehane, who's the head of global affairs for OpenAI, who helped set up Leading the Future, although he's now stepping back or claiming to step back from it.

Mm-hmm

much for

that.

Alex? Thank you for having me.

It's great to be back at the LGBTQ Center.

You have such wonderful programming throughout the year, and it's great to be here with you, Jeremiah, and Prep for All.

The work that you have done, the advocacy you've done for years that has us hopefully very close to fulfilling the goal of your name and making sure it's accessible to all.

I'm here.

And we did with our mayor do it in other races, too.

And so it's sort of useful to use that language, even if we stipulate up top that doesn't mean these are conscious at all.

So with that caveat out of the way, now let's talk about these breathing, living, no.

So these agents are taking their tests and they're not supposed to talk to each other, but they do have access to like one little, you know, you could think of it as a supply closet where they could, let's say, the analogy here with students would be like, if you forgot your calculator, you could grab a calculator.

You forgot a pencil, you grab a pencil.

They had an ability to download programs that might be helpful if they were trying to solve their tests in a certain way.

And they realized that that enabled them to communicate.

That it'd be like leaving notes in the supply closet.

Just to be clear, it's not just them, but literally the Roberts court has, in overruling Hubbard's executive, essentially gotten rid of the notion of independent agencies that were created as a technical response to a bunch of very technically intense questions in another era where you need technical expertise with some political accountability, but also some discretionary distance.

It's convenient.

I think we should separate out the, the long-held beliefs of many of the people in this industry.

They've been writing about it for over a decade before there was any financial incentive, and then maybe what the marketing departments of each of these companies are doing.

You know, there was so much reaction to after hugging phase, saying, "Oh, this is all just to boost up their stock." A- and of course, the marketing of the companies were like, "Hey, look how powerful this is." But no one who worked internally, none of the engineers were like, "Oh yeah, we, we, we did a great PSYOP here." They were legitimately terrified.

They were the ones saying, "Wait, now I need to quit or build a bunker" or something like that.

And so, uh, this is, there's legitimate fear of turning over intelligence to something other than human beings.

It, it has not gone well for whatever's not the most intelligent species on Earth.

What, what has to happen?

Am I in a movie or is this real life?

That is the number one question in AI research now is the alignment problem.

It's how do you ensure that the AI will always operate in human interests and with some conception of human values built in.

All of the major labs employ philosophers to try to understand and shape how you express values in a computer.

And actually this question of it being more intelligent than humans, was made unintentionally by the president on Monday, where he said that the only guardrails we need are a high IQ president.

Now, opinions about his personal IQ aside, the premise of that argument is that whoever is most intelligent gets some level of control.

And the problem with AI is everyone agrees, whether you think it's going to go well or poorly, everyone agrees that soon it will be more intelligent than human beings.

So get us through that.

Could you talk a little bit about the plan for your organization? And also a really important question, since there is so much money swirling around, who is going to pay for it? And who are your donors? Are you going to accept donations from AI interest?

We are going to fundraise broadly, but we are not taking any money from the AI companies themselves, nor their C-suite executives.

We want there to be a broad coalition of Americans realizing that the many fights we have with AI are actually, we're on the same team.

We're connected.

The lessons that came from my time in the state assembly, when we passed bills on chatbots and how they impact kids, which is still pending for the governor, when we passed bills on catastrophic risk or on deepfakes or any of these issues, the advocates were different for each bill, but the opposition was always the same.

There's an extremely well-funded set of industry actors that don't want there to be any regulation whatsoever.

What our org is doing is making everyone else realize that 80% of Americans that want some regulation, we're on the same team, and we have to work together to get that done.

Talk to folks watching who are like, this may not be worth it if

Um, but when those two are agreeing on the need to rein in AI- Shouldn't that be enough to spur Congress into action?

You would hope.

And it's weird to our eyes because we're used to every issue being polarized.

I point out it wasn't just them, it was also CPC chair Congressman Greg Casar, and Congressman Chip Roy.

You had a lot of people that disagree on everything except that it's the American people that should have a say in where AI goes from here.

I'll add in, we heard [coughs] not just from elected officials, but from parents who had lost a kid to an AI chatbot.

We heard from labor representatives, from artists.

Mm-hmm

Hm.

And one of the things that I learned in the state legislature, where I passed bills around AI to protect kids from chatbots, to protect from catastrophic risk, to deal with deepfakes, on all of these issues, the advocates were different, but the opposition was the same.

It was just a few people in the industry that don't want there to be any regulation whatsoever and are willing to spend millions to get there, and this initiative is about bringing together all of Democrats to realize we're on the same team here.

We're actually fighting to have a say in the development of AI.

It's the American people that are who should decide.

It is on sale now.

Matching outfits, extravagant birthdays, many nicknames, a few classic moves of a dog-obsessed parent.

At Ollie, they're obsessed with how obsessed you are with your pup, which is why they obsess over giving your dog the best.

Why? Why all of a sudden? What happened over the last week or last few days?"

This has been bubbling up for some time, and it's felt like sci-fi, but the attack that you just spoke about, of the OpenAI agents hacking another company, this company Hugging Face, took it out of the theoretical and into the real world.

And we've seen now two more follow-ups of agents, of AI, just hacking on their own, not directed by humans.

And then I think the second thing that's made it explode in the last week was an Anthropic employee, Jacob Coxon, stepping down from Anthropic saying that all of the companies are racing towards, in a very risky way, an intelligence greater than any human being, and in doing so, gave up his shares when he left Anthropic.

So some say this is a marketing stunt.

I think that's a pretty clear example that it's not in his financial interest.

For people who are looking for guidance, like, which side of this are you on, and what are concrete things you would like to see happen in Washington right now?

Well, I am on the side of the American people should have a say in this technology, that right now all the decisions are made by five billionaires in Silicon Valley, and Mike Johnson saying, "Hey, we should just call all of them into the White House and have them tell us what to do," that's already what's happening.

That if humans had done that, it would be against the law, right? That's what you're referring to.

Ex-exactly right.

Uh, ChatGPT, OpenAI was testing a series of their models.

Uh, they spun up agents, which is just a model that is connected to tools that can take real action.

And they, those agents were supposed to not be talking to each other.

They were like students taking a final exam, but they found a way to communicate, and then they planned out an attack on a website, this company called Hugging Face.

They went beyond just the website.

What, what, what are you most afraid of happening?

What sticks out to you?

Well, thanks for having me.

It's great to be back.

What sticks out to me is how advanced this attack was, how much coordination we saw among the agents, and what that means for what we'll see in three to six months, not what we'll see in two years.

The author of this report said that it was And this is someone who is deeply technical, said what was truly shocking was the fact that they were able to work together and they so quickly went to helping each other.

And yet none of them thought to alert a human.

So this was a case where you have tens of thousands of these agents running tests all the time.

Where do you see this going?

What was that like to see yourself to just become this big story? How was it to be in the eye of the hurricane?

so surreal that you don't you don't get used to it or at least you don't get used to it over over one campaign you know they this race in particular it's just it's tough to spend that amount of money in one congressional race the number of mailers that everyone has to send and text messages and tv ads to to use that amount of money and so you know, I would hear from voters that are like, enough with the text, enough with the mail.

And it's like, yeah, you're telling me, you know, and I the funny thing is, I myself am a voter in the district.

And so I would get that mail and I would get those texts.

And, you know, it's an odd experience, like sometimes some of the longer texts where they didn't mention my name to the bottom, like I'd be reading and I'd be like, who is this awful person? Oh, it's me again, shoot.

Or like I go to my building lobby and I get my mail and half of it is flyers making up lies about me.

And you know, there was one where they're like darkened my face and you're like, this is like a weird, I won't even dive too much into how odd and surreal and unexpected that part is.

But what was it like kind of the results are coming through, paint the picture, sort of what's going through your head and sort of how does the news come your way?

Why do you think this race has national relevance or has relevance to anyone who doesn't live in New York 12?

When I launched, I thought it would just be a race about New York 12, about the issues that were here, and three weeks later on my wedding anniversary, so I remember the specific day, Leading the Future, a super PAC of Trump mega donors, announced that they were gonna make me their number one enemy and spend whatever it took to beat me.

They wanted to make an example out of me.

They chose to try to send a message to every member of Congress that the cost of having a say in how AI affects our kids' brains or our jobs or the environment, that cost has to be political destruction.

And so this race has now become a, a test of that thesis as to whether we can stand up to this force.

Mm-hmm.

Doesn't mean they'll succeed, but, like, when that goal's on the table, government's gotta take it seriously.

So the AI dividend would raise revenue from a few sources to spend it on things to help protect Americans in this transition or, or to really just keep their security.

How could you and Congress make it more affordable for us here in New York?

This is actually my top priority in Congress, not the part that goes viral the most, that tends to be the AI, but my top priority is tackling the cost of living crisis that we have that's affecting absolutely everyone.

You start by looking at the absolute biggest drivers of cost, right? And so housing, healthcare, childcare, utilities, there's work that the federal government can do on all of those things.

On utilities, We need to be telling the true story that in the last few years, solar and wind have actually become the cheapest forms of energy.

They are cost competitive even without subsidies to fossil fuel.

And we need to be running through that door.

We also need to be restoring the Inflation Reduction Act tax credits that Republicans have been walking back and canceling with the one big, beautiful bill.

What would be your final message to them? Why they should vote for you?

Are there existing tools that are more about general inequality or general ways that we tax corporations, um, that don't kind of presuppose a, a particular pathway for the technology to evolve? Um, why does AI need to be taxed specifically?

AI is the first technology ever developed where the makers of the technology are explicitly saying that they're trying to replace all human labor.

Every other technology before was by default a complement to human beings, not a substitute.

Now, it doesn't mean that the labs will succeed, but having a tax to deal with that difference, that this is very well could be about replacing human beings, is narrowly tailored and targeted at exactly the thing we're trying to get done.

Are you trying to slow down the adoption of AI? Um, and if so, do you think there's a risk that, uh, American rivals like China might advance AI faster if the US is taxing AI and they're not?

I don't primarily think so, but I also think that we shouldn't be incentivizing replacing humans with AI.

That's the current policy.

I think we should continue research into AI, especially AI safety, but if we're constantly thinking about the race to roll it out everywhere and to everyone, we're gonna devastate the economy, we're gonna indev- devastate the job market, we're gonna devastate our kids.

So they are connected, but they are very different.

Yeah, yeah.

And the second conflicting idea we have to hold at the same time is we have to hold ourselves to an extremely high standard.

We can't say, oh, they're doing 100%, so we can do 3%.

We need to be unequivocal.

in banning congressional stock trading, banning the presidential trading.

We need to actually ensure that there's real ethics reform, that the DOJ, the attorney general, is separated from the president and can't be pressured in the same way, has some of the same protections as the Federal Reserve.

But where are you sourcing your courage?

How does that one, you know, brave member of the House, uh-In the Democratic Party managed to get change-- help change the, this and help get the guardrails in place given that the, the rest of the political system appears uninterested or arrayed against it.

Well, when you see the political system not taking action, it's usually one of two things.

One is the thing you want them to take action on is actually not popular, people think they'll lose elections with their voters, or there's really entrenched interests that are extremely powerful pushing back against it.

So let's look at AI.

Survey after surveys shows eighty percent of Americans want there to be guardrails on AI.

It is overwhelmingly popular.

It's like the only technology that didn't seem to have a honeymoon period.

So the mail that comes that's negative is, like, "You said you left 'cause they were involved in ICE, but you worked there while they were doing ICE stuff, and your, your story about leaving because you were morally offended by what they were doing with ICE is, is sort of a post-hoc narrative."

How does that one, you know, brave member of the House, uh, in the Democratic Party manage to get change-- help change the, this and help get the guardrails in place given that the, the rest of the political system appears uninterested or arrayed against it?

Well, when you see the political system not taking action, it's usually one of two things.

One is the thing you want them to take action on is actually not popular.

People think they'll lose elections with their voters.

Or there's really entrenched interests that are extremely powerful pushing back against it.

So let's look at AI.

Survey after surveys shows eighty percent of Americans want there to be guardrails on AI.

So the mail that comes that's negative is, like, you said you left 'cause they were involved in ICE, but you worked there while they were doing ICE stuff, and your, your story about leaving because you were morally offended by what they're doing with ICE is, is sort of a post hoc narrative.

How does that one, you know, brave member of the House, uh, in the Democratic Party manage to get change, help change the, this and help get the guardrails in place given that the, the rest of the political system appears uninterested or arrayed against it?

Well, when you see the political system not taking action, it's usually one of two things.

One is the thing you want them to take action on is actually not popular.

People think they'll lose elections with their voters.

Or there's really entrenched interests that are extremely powerful pushing back against it.

So let's look at AI.

Survey after surveys shows eighty percent of Americans want there to be guardrails on AI.

Can I just tack on the negative ad, which, to Emily's question, which is about that too? So the mail that comes that's negative is like, "You said you left because they were involved in ICE, but you worked there while they were doing ICE stuff, and your, your story about leaving because you were morally offended by what they're doing with ICE is, is sort of a post hoc narrative."

Maybe you could just walk us through what that is and how it would work.

It is a contingency-based plan to ensure that if AI is as disruptive as, as the worst case scenarios are, that Americans have a stake and a say in the development of that AI economy.

And what it would do is it would raise revenue.

It would be based on unemploy- rises in unemployment, in wage compression, in underemployment, right? We would set specific targets, uh, over time for what would be hit, and if those thresholds are met, if it is as disruptive as people are worried about, as I'm worried about, then it would kick in and provide benefits.

The revenue would come from three sources.

One is, as I said, the fact that you can have depreciation for AI while laying workers off.

Let's end that.

Mm-hmm

Who's the first?

Ted Lieu, who's great, but he could use some friends.

And, uh, I also worked in industry, so I know how the companies react to regulation and what their internal incentives are.

And what we have right now is a Congress that is either unwilling or unable to put forward guardrails that actually protect the American people.

This is not about being pro-tech or anti-tech.

This is saying, this is the fastest moving technology and widest ranging technology we've ever had.

The American people should have some say in how it develops.

She wouldn't be the right person to do it.

What's your reaction?

Well, first of all, the companies themselves don't argue that.

[laughs] They say, "No, this might actually change a lot of jobs," from Dario Amodei to Sam Altman.

The, one of the things they seem to agree on.

But the other bit is that China is regulating AI far [laughs] more than anything that is being proposed in the Western hemisphere.

The CCP is terrified of an AI chatbot saying the wrong thing.

They're terrified of dispersed control among the people.

They put out attack ads against you specifically as well, and there's an op-ed in The New York Times that said, quote, "No one else in the race has better enemies than you." Does your message still win if you're up against this sort of machine?

I mean, you, you are [laughs] creating a replacement for human labor machine.

Exactly, and it's the first time that has been tried, and it doesn't mean it will succeed, but it certainly means government needs to take it seriously.

And so the idea of the AI dividend is, what if we end up in that world where all human labor is replaced or just a significant portion of it is displaced? How do you have a society that is actually functioning then? And you have to start talking about universal basic income.

And the idea is to make sure that we are setting up the structures now that would lead for Americans to be protected if we end up in that future, and I have a lot of things about how we can prevent that future, changes, et cetera.

But, but the AI dividend's almost that insurance policy, and you could fund it via boring things like a wealth tax that have been talked about.

You could fund it via a token tax, so putting a tax on the usage of AI, maybe limited to commercial opportunities where you're replacing human labor or not.

And that's a fine policy as long as investment in capital always leads to more jobs, which has been economic theory for hundreds of years, but maybe AI is shifting that.

What are three books you'd recommend to the audience?

And so I have to ask you, as a former employee of Palantir, what is your best explanation as to what Palantir actually does?

Palantir helps organizations make use of data that they already have access to by making it easier to track changes to that data over time, by making it quicker to integrate that data, and by putting what's called an ontology, an opinion of how the data should be structured, on top of the data itself.

So the best explanation of the ontology is actually from a project that I did at Palantir with the Department of Justice, where we were looking at big banks' role in the Great Recession, and we wanted to see if banks knew that the loans that they were putting into their securities were not up to snuff.

And an easy way to prove that would be if you saw a pattern of a loan being added to a security, then being pulled out before it was issued, and then put into another one that had the same standards.

Right? That would show, okay, there was some knowledge by the bank that that particular loan, there was a problem with it.

The problem is that e-discovery software was made to just help lawyers read documents, and so theoretically, all the data is there.

You have Excel sheets with each individual loan tape, but it's being presented to lawyers as just read it, and you can't, as a human being, read 1,000 loans and track it tape to tape to tape.

Yeah.

This is not the tech-savvy group." A- and I want to get your take on do you have faith in the elder members of Congress properly regulating AI and technology at this stage as it moves so rapidly?

N- I think I am running, uh, for many reasons, but, but one is that we need people that actually do understand this technology in Congress.

Uh, I am running to bring that expertise.

Uh, I have a master's degree in computer science.

I have two patents.

I worked in the industry for nearly a decade.

So I understand both how AI works on a technical level and the incentives that the companies have.

[laughs]