Assessing the Impact of the Midterm Elections on US Tech Policy

Audio of this conversation is available via your favorite podcast service.

The United States midterm election season proved that tech policy is no longer a side issue in American elections. AI, data centers and the political power of Big Tech and oligarchs like Elon Musk are now things candidates have to take a position on — and voters are paying close attention.

After the polls closed on Tuesday, Democrats posted sweeping wins up and down the ballot. If their victory speeches are any indication, efforts to hold Big Tech accountable and to address data centers and AI governance will continue to be defining issues in Washington and in state capitols across the country.

To unpack what the election results could mean for the future of US tech policy, we pulled together a live podcast discussion with guests including Alexandra Reeve Givens, CEO of the Center for Democracy & Technology; Robert Weissman, co-president of Public Citizen; and Brian J. Chen, policy director at Data & Society.

What follows is a lightly edited transcript of the discussion.

Justin Hendrix:

I am pleased to be joined today for this conversation on assessing the impact of the midterm elections on US tech policy. I've got an excellent set of panelists that are joining us. I've got Brian Chen, who's policy director at Data and Society, Alexandra Reeve Givens, who is the CEO at Center for Democracy and Technology, and Robert Weissman, who is co-president at Public Citizen. These folks have a lot of experience obviously working on tech policy issues, working on Capitol Hill, and thinking about various issues that sort of intersect with tech policy, including corporate power, including the way that Congress works. So I'm hoping that we can have a good discussion today and maybe push a little bit past the headlines that we've read just to date in the last 48 hours or so, see where these folks are thinking and maybe get a good sense of where we ought to be focusing our attention over the next few weeks.

So I'm going to just start things off, hopefully giving everybody a chance to give a general takeaway, something that they are thinking about that's sort of top of mind for them. And I suppose the way I might want to start is by asking what you think will endure about the results on Tuesday. I mean, clearly there's been a lot of headlines. We've seen a lot of sweeping coverage of a degree and the scale of the Democrats' victory across the country, both in Congress but also across the states. But I'm wondering what you think will endure, what we'll still be talking about a year from now about the effect of Tuesday's elections. And Rob, I think I might start with you on that. I mean, your organization's talked a lot about what this election says about corporate power, corporate spending, corporate influence on our politics. What will endure a year from now?

Robert Weissman:

Yeah, I think that some of the big themes are going to be generic, but also very specific to the tech sector and you were beginning to touch on them. So we saw this is the most expensive election in American history and we saw by far the most direct corporate spending in American history. We saw huge investments by AI and big tech companies overwhelmingly on the Republican side. We saw Elon Musk put on, we don't know yet how much, might've been up to half a billion dollars in this election, and that was a big fail. And we also saw in races around the country pretty intense resentment toward big tech, toward AI, and especially on data centers. So I think as we move forward, there's going to be in Congress, and we'll see about the administration, but in Congress, a shift towards greater demand for corporate accountability and a new willingness to go after big tech.

We've really seen things stall in Congress even after an initial moment of ChatGPT 3.0 where it's like maybe there'll be regulation that lasted about six months and it's gone. Now we may have another window where that's going to happen, but it's going to be very contested. So without going on too long and filibustering, maybe that's a big topic. I think there's an opening, but where it's going to go is very much up for grabs.

Justin Hendrix:

Alex, how about you? What will endure a year from now?

Alexandra Reeve Givens:

Yeah, I mean I think people are talking about this as the affordability election, that the thing that really was driving people is just how difficult it is to get by in the course of their daily lives. But when you pull back the broader theme around that I think is people feeling deeply disempowered and both the economy and the government aren't working for them. And so that I think is going to be a theme that endures in terms of the priorities that Congress has to pursue, the language and framing that both members of Congress and state legislators and advocates will be using to help show the salience of these issues to people's regular lives. And that ends up meaning a lot for the advocacy agenda. And also just this question of how society navigates when we continue to hear about the massive transformation that is going to be shaping our country and our world and lawmakers sit idly by without engaging on it or providing meaningful guardrails, the American people want action.

And so both, I think this grounding of affordability, wealth questions, lack of trust in institutions and lack of trust in the companies coupled with a real desire to see governments hearing those concerns and taking concrete actions to address them.

Justin Hendrix:

Brian, how about you?

Brian J. Chen:

Yeah, I mean in terms of what endures, probably the just sort of election as pushback against Trump in general is the most obviously the thing and the force, the political force that's going to endure the most. That is going to be the most resilient, that is going to have the most profound implications for our politics, for the GOP party, for the Democrats. I think as far as tech, I might actually say there's certainly the popular opinion is very clear. People are very anti-AI, they're very anti-data center. It is a little bit more tricky I think to say that endures indefinitely because it's going to go through some... There are going to be really interesting effects as we start to see that go from this protest vote against AI and data centers and translating that into an actual governing agenda for what politicians want to do about that.

Clearly there has been a mandate, people don't like these data centers in their backyard. They don't even like them in their neighbor's backyard even if they live quite far away. They don't like the effects of AI in education in all kinds of parts of society. But in terms of how are you actually going to govern and deliver on that for politicians, one thing that I've been paying attention to is just thinking about every single politician has... It's not even partisan at this point. Every politician says we are against AI, we want to do some restrictions on data centers. But now that there's an actual governing moment, we're going to have progressives and centrists and conservatives come out with a variety of policies and they're going to say, "We are sticking it to big tech. We are doing some kind of restrictions on data centers. We're doing restrictions on the use of AI." And it is an open question to me.

I think which of any of those is going to matter the most to voters? Do they want the bread and butter stuff like affordability, making sure their utility bills don't go up? Do they want broader anti-oligarchic measures for the economy writ large? That's the stuff that I think is more sort of amorphous.

Justin Hendrix:

You've been writing about this, Brian, in particular, you get a piece in Dissent that focused on the politics of data centers. Are there particular races that you're looking at around the country that indicate to you which way we might go on those questions?

Brian J. Chen:

Well, one state that I was looking at quite a lot, and I want to shout out the work of my colleagues at Data and Society, they put out this report on Pennsylvania, The AI Factory that was really just this great deep dive on the AI build outs in that state. Pennsylvania was interesting because there, the governor's race, Josh Shapiro won big. There was no drama about whether or not Josh Shapiro was going to win, but he really wasn't trying to win Pennsylvania. He was trying to position himself as a national leader for 2028. And so what Shapiro did with respect to AI and data centers is, I think, quite illuminating. How he moved during the election is, it tells us something about how politicians want to triangulate their position as a reaction to this opposition, this popular opposition to big tech. And so we saw, and as my colleagues documented, early in the cycle, Shapiro was a huge data center booster.

He really wasn't that far away from the messaging you might hear from President Trump or a representative of the tech industry. He talked up about the generational opportunities of this private investment in the AI data center construction. He wanted to fast track projects to invite development in Pennsylvania, and he had lines where it was like, "This is really important for national security, and this is important that we lead in innovation and that we beat China in innovation, and we're going to create all these great construction jobs." And as people in Pennsylvania became further organized and mobilized and voiced their opposition, you saw a pivot by the governor and the governor's office to say, "Okay, data centers are important, but we recognize what people are saying. We need to do some slowdowns. We need companies to bring their own energy. We want there to be more community engagement." And then much, much later in the campaign, as the data center sentiment just really, really worsened over the summer, Shapiro released an executive order that really doubled down on what his office called the strictest sort of restrictions on data centers, focusing on energy affordability, on environmental protection.

I think the question is, did the governor actually change his policy position or did he just change the political messaging? There have been a number of organizing groups in Pennsylvania who said that even though Shapiro calls these the strictest limitations on data centers, there's sort of glorified restatements of what the law actually says already about the permitting process in the state. And so politicians, I think, recognize the populist politics. They understand that it is a winning message to be against data centers, and yet there still is a brand of politician, powerful politician that is synthesizing that with this political strategy of you need to hold together this coalition of voters and organized labor and capital. And so how much of that anti-data center resentment really, does it turn into just milquetoast protections or is there something more transformative, more powerful that politicians can offer as an alternative?

Justin Hendrix:

Rob, Alex, I don't know if you want to jump in on this and if there are maybe other exemplar races you'd want to point at.

Robert Weissman:

It's amazing how data centers were an issue really across the country. If you happen to be traveling and you popped into a state and turn on the TV, there it is, there's an ad on data centers. But I think too where it was central were Ohio and Texas. Ohio, Sherrod Brown, he made data centers and labeling his opponent as data center Husted as central to the campaign and was able to turn all the support that his opponent had been getting from the crypto industry and from AI against him pretty effectively. And in Texas, it was also a central issue. And there, I think it's very interesting what happened in the governor's race as well as the Senate race. In the governor's race, that's the discussion we just had about Shapiro in Pennsylvania. In Texas, Greg Abbott, obviously to the right of Josh Shapiro, pivoted so totally that he imposed a statewide moratorium on new data centers.

No new permits were going to be granted for any data center construction. It happens we have an office in Texas and we found that while the permitting agency was still granting permits because they actually literally have them as a rubber stamp. They're automatic. You apply and they're automatically given, no one even looks at it. So we called it out and then the governor intervened and they shut down the permitting. So if you saw that happening even in Texas, I agree with Brian completely on the overarching point that there's going to be a big fight about how much this sticks and particularly among the Democrats and to some extent the Republicans, how much they're accepting the populist moment or how much they're still bowing down to the donor class. That said, I think the data center issue, which is an issue in its own right and also connected and sort of a signifier for the broader question about big tech oligarchic power, it is real.

And if it was real enough to shift the position of Greg Abbott, no politician is immune.

Alexandra Reeve Givens:

Yeah, and Rob landed right where I was going to go, which is just one of the things that is remarkable to me is in every single election, whether it was for Congress at state level as well, we see people leaning into tech policy issues and it wasn't just data centers and it's making it into the victory speeches, it's making it into the committee chairmanship memoranda that are circulating this week as people make their bids for chairs of different committees. And so you really are seeing a focus there. And time and again in each of those instances, again, whether it's data centers or it is about privacy AI governance, it is about this question of whether or not companies and the government are responsive to Americans' daily lives and the issues that are impacting them. And so that moment of tech issues really hitting kitchen table issues, which many of us who have worked in this field and Justin, you curate the newspaper for our field or the blog for our field, we've known this for a long time, but seeing it pierce through to this level of dialogue I think is a really important shift.

Justin Hendrix:

Yeah, maybe just stick with that just for a second because it feels to me like that's the thing that's really changed is that tech policy is now the everyday kitchen table issue on some level. The tech politics has gotten very, very interesting. I feel like we're seeing people connect issues now in their head. They're seeing AI in schools, social media and their children's lives. They're connecting that to concerns over surveillance, concerns over data centers, concerns over jobs and artificial intelligence. It seems to me that, at least as far as the rhetoric's concerned, we did see Democrats seem to make that connection relatively successfully, but it strikes me it's a much harder thing to address. I don't know, what do you think needs to happen to address that, Alex, in the next few months to satiate that desire to address those politics?

Alexandra Reeve Givens:

I mean, I think we're going to be looking at a couple different tools in the toolkit for members of Congress, and one of them is going to be an active oversight and investigations agenda, and we've seen that even in the course of the past 48 hours in social media rallying cries around subpoenas. Then of course there's going to be pushing the legislative agenda, which to be clear, even if bills aren't crossing the finish line, we're in the race for 2028 now. And so members are going to be showing what they stand for, what the solution space looks like and showcasing their priorities through the legislative pushes that they make. In terms of the oversight agenda, this can go in two different directions. So one, and I imagine Rob will talk more about this, will be looking at some of the company practices that are impacting Americans, particularly as pocketbook issues, kitchen table issues, and doing oversight where the government so far they believe has been falling short.

But the other that I really think and hope we spend time with is oversight of the unbelievable attack on constitutional rights and values that we've seen over the past two years and the real concerns about corruption and grift and the targeting of people across the country. And one of the challenges, I think it's both an opportunity and a challenge. The challenge will be spending enough time on those issues to do them justice, to create the marker point in history of these being unforgivable oversteps that we should not repeat, and also delivering on that agenda of showing the American people that you are addressing kitchen table issues and affordability. But where those two come together is these concerns around corruption and abuse of power. I think if the Democrats, what I anticipate watching them do is tie those two things together and thread the needle and be able to have both the democratic reckoning that we need to help ensure a healthy democracy and robust safeguards moving forward and happy to talk about what some of those issues are and also really speaking to this moment of tech governance that's needed as well.

Brian J. Chen:

And just to jump on Alex's point, I mean, we all remember that photo during Trump's second inauguration where all the tech leaders and CEOs are kind of front row at the inauguration, and we take for granted that that's really weird. Why was that the case? Why did that happen? It really underscores Alex's point that there is a level of corruption and self-dealing and coziness with the tech industry that I think would be smart politics to go after in terms of using oversight, doing investigations with your new House majority or Senate majority. There's really no other place to start. I mean, there's many places to start, but one of the ones that absolutely has to be on that list has to be DOGE. I mean, that is such a clear example of abuse where the White House was partnering with a tech trillionaire, Elon Musk, maybe today he's a billionaire at this point, inviting all these sort of quasi-contractors in from Silicon Valley to go through agencies to figure out where is the fraud or abuse, so to speak, and really just laying off people who have been career civil servants.

That chapter really needs to be resurfaced and Democrats need to be taking a look at that and say, "What actually was happening there? What was the role of the tech companies there? What was the back channeling between these tech leaders and senior officials in the Trump administration to facilitate this sort of DOGE process?" Because it really speaks so much to this fusion, this corrupt fusion between the state and industry.

Robert Weissman:

When I checked yesterday, he was down 850 billion. So today maybe it's back up, I don't know. The limit is going to be the number of days in the year in terms of the oversight agenda. The policy agenda is a little bit, I think, more knotty and complicated where it's going to go, but I think we should anticipate seeing oversight related to tech policy across committee sectors. And maybe we've just highlighted three areas, which I think is right. One is going to be just the naked corruption and the big tech embrace of the Trump administration, and all that means all the ways that money has been funneled to Trump and his related things through the ballroom inaugural committee, these corrupt settlements of fake lawsuits by Trump, et cetera, and then what they're getting back in exchange. Two is going to be, I think, the governmental policy towards tech, towards Musk specifically, and number one, but really widely.

There's a lot obviously to be looked at and the sort of aggressive combination of laissez-faire and also simultaneously state industrial policy and critical minerals and more just to kind of a morass of secrecy and corruption. And then last, because there hasn't been any federal regulation at all, which in fairness is not only a problem of the Trump administration, but is a specific problem of the Trump administration, because Alex was saying they're the practices of the companies themselves, including I think some of the frontier AI issues and the hacking questions, hacking is sort of the understatement of the problem, and also I think a lot on focus on companion chatbots and impact on kids. I guess I skipped over, I just don't want to ignore the point that Alex made in the governmental policy space, the issue of surveillance and abuse of civil liberties and the facilitation of the cruelty and brutality of ICE by various tech companies I think is also going to appropriately and necessarily come under close scrutiny.

Justin Hendrix:

Well, let's spend just a minute maybe laying out for our viewers and listeners what we can expect in different committees or what you anticipate, at least on what you're seeing right now. I mean, we know that the Democrats apparently are, it's been reported planning a kind of multi-pronged effort to look into Musk in particular, and it seems like that's both necessary given the scale of some of the things that you've described and that Brian described in particular, but also somewhat likely popular in a way that addressing Musk as a kind of singular exemplar of a lot of the corruption and the problems that we're seeing around tech and capitalism in this country makes a lot of sense. He's also very close to a lot of the concerns over AI and national security and influence abroad, all of those things. But what else can we expect?

What committees do we expect to be particularly active, both investigative and oversight? I assume, Alex, if I remember correctly, you were on a Judiciary as a staffer. I assume we'll see a very busy judiciary committee. What other committees across both the House and the Senate can we anticipate being very active?

Alexandra Reeve Givens:

I mean, I expect it's going to be a mix and that's probably a good thing to be able to staff this appropriately and spend time. I'd love to talk about four areas that I think are going to be really important for oversight, and I'm going to focus on oversight of the Trump administration and then know that we can turn later to more commercial practices, some of the other things that Rob was listing in his taxonomy. The DOGE one is really important, and obviously there's the Musk angle that I think helps frame this in the people's mind in terms of there being an avatar to represent what that looked like, but really drilling down on what was an extraordinary illegal grabbing and consolidation of Americans' personal data by the government, trying to make sure that that doesn't get lost in a wonky discussion, and instead people realize the real assault on constitutional principles and values and the American people's trust that embodies, right?

So we have DOGE being able to access and consolidate people's highly sensitive data from the IRS, from social security, using it for immigration enforcement, federal agencies demanding that states hand over deeply personal data about SNAP and Medicaid recipients. They're breaking laws, they're being sued for this, their injunctions being issued. When people do get their hands on the data, you have random DOGE staffers getting read-write access over the treasury's payment system, evidence of them using personal laptops. So just this very flagrant disregard for the importance of these things, and the fact that these are tied to social safety net programs and programs that people when their beneficiaries really do hold dear, I think is an important element to unpack, not least because it needs to lead to legislative reforms, but the oversight piece of this is essential. So that's one. Two is going back to this surveillance question.

So how do we create this national reckoning on the assault on Americans' personal privacy and the right to protest? So we've had examples of ICE treating dissent as criminal conduct, following protestors home and sharing personal facts about them to intimidate them from protesting, using face recognition in the field, threatening protestors with arrest for filming ICE agents, monitoring and mining dissenter social media feeds to be able to identify them, and then of course all the ideological-based screening being used in immigration decisions writ large, so reporters being banned access to the country if they've covered a protest before. So that is a really important vein, and we've seen members of Congress talk already just this week around what the oversight agenda and what appropriate controls will look like there. Two more that I'd highlight. The next is this assault on the First Amendment that we've seen, including the ideological takeover of the FCC and the FTC.

Two agencies, which to be clear, should be defending American consumers and should have that as their day job every single day at a time when our country is undergoing enormous technological change. Instead, what we've seen is these absurd unconstitutional efforts to get people kicked off the air because you don't like their coverage or revoke a broadcast license because you think that a program is too woke or conditioning a merger approval on somebody dropping their advertising, changing their advertising policies in the Omnicom case, hearings on gender-affirming care and it being a deceptive trade practice. I mean, time and time again of these chairmen of these now deeply political agencies because the Democratic commissioners were fired, very clearly pursuing a political agenda, in my opinion, deeply at the expense of the things that they should actually be focusing on to serve the American people. And so making sure that those abuses are highlighted, and again, tying it to this question of what self-serving ideological work looks like versus addressing Americans' everyday needs.

And then the fourth one that I'll touch on, CDT has an elections and democracy team. We are deeply focused on the efficient and fair functioning of elections. Luckily, this election did not see major crises or discrediting, but we have to be really clear. It was not for lack of trying by the administration to significantly shake the foundations of our electoral democracy. Lawsuits against the states demanding that they hand over detailed voter information, the administration trying to wreck mail-in voting, impacting a third of voters across the country who vote by mail, attempts to sequester mail ballots, ballots, threats of prosecuting election officials, and that's just been the past couple weeks. There, again, we have to have the reckoning, number one, for accountability and to make sure that we are placing that marker in history, but two, to shore up our defenses for the work and time ahead.

So there are many other topics that we can cover, but those four alone are really big and really important. They could each fit into different committees, so we'll let the majority leader figure that one out. But I do think that each of those is going to be really important to help make sure that we're making progress and holding the administration to account for the abuses that we've seen in those spaces.

Justin Hendrix:

Rob, Brian?

Brian J. Chen:

What I appreciate about Alex's length of all those oversight examples is each of those is really constructive. There's versions of oversight where people are twiddling their thumbs and doing a lot but not actually accomplishing anything, and it's just sort of like, "Well, Trump's a bad guy. Elon Musk is a bad guy." All of this stuff that Alex pointed out, it's sort of like this is how your rights have been affected. This is how your information and data has been impacted. This is how your benefits have been made more inaccessible. All of this is sort of really helpful as constructive politics if we think about oversight along this two-year window that we're trying to build a narrative, build a longer term vision that hopefully can transform into something in the 2028 election. So I just wanted to add that little piece to it.

Robert Weissman:

Yeah, I would say from an advocacy point of view, we think that the point that both Alex and Brian are making is super important that the oversight... There's tons of opportunities for gotcha moments, and that's good for individual social media accounts and stuff, and totally useless for shifting the political landscape. So we're really encouraging that the oversight be very intentional about setting up the policy reforms that are the remedy to the abuses that are being focused. On the broader question you're asking, Justin, I would say that I'm not sure what you're going to do with the name of your publication. I think it's going to be hard for tech policy to be thought of as a discrete thing if it still is, because these are the largest companies in the history of the world with this technology, which if it's a hundredth as grand as they claim it is, is going to be transformative and economy-wide.

So I think we're going to see things that are called tech policy or things that aren't called tech policy where tech is sort of the dominant piece of what we're talking about across committees. So in the tax committees, for example, they're going to have to deal with corporate tax. They're going to have to deal with whether they're special taxes on AI companies or the corporate tax issue directly, and specifically the problem of the big tech companies evading the base level of taxation, which they're masters of doing. We're going to talk about antitrust policy and anti-monopoly policy, and that covers all sectors of the economy, but it number one, focuses on the largest companies in the world, and those are all the tech companies. In the national security space, you're going to have, on the one hand, I think it'd be a lot of focus on what Musk has done in Ukraine and what it means to have a person, not even about a company, but a person in charge of this critical global infrastructure.

But also we're going to see huge questions also involving Musk by many others about AI enabled weaponry. And they've even created a commission on the so-called Future of War, which is basically an effort to expedite the advancement of AI enabled weaponry. So it's going to be present there. And the education and labor workforces, we're going to see a big focus on trying to get ahead of what is AI going to mean for the job landscape. And if it's nearly as disruptive as the companies themselves claim, what are the remedial measures to soften the blows and ensure transition and so on? In education, we're going to see it. In healthcare, we're going to see it. It's really going to be pervasive, I think, in the oversight agenda. Whether or not people think I'm going after the big tech companies, whatever they're going to be focusing on, given the role of these companies and the technology in our society now, it's going to eventually come back and include them.

Justin Hendrix:

What prospects do you all see for any specific legislative gains in the next couple of years? I mean, it strikes me that so many of the basics are still unaddressed, privacy probably being one of the biggest ones. Is there any prospect you think for some of these things to maybe cross divided Congress, maybe even cross the president's desk in the next two years?

Alexandra Reeve Givens:

I view it as part of my job description to be the person that always says this is the time for privacy. And I do it with full knowledge about just how difficult it is and yet we need to fight. I do think there are a couple of things that change the landscape, and this is true both on privacy and AI governance. On the privacy piece of it, this remains a bipartisan issue. Voters across the country care about this question. AI is revealing every single day how great the need is for greater regulation and greater controls for the enormous amount of information that we share about ourselves in the course of our daily lives using these chatbots, connecting these agents, and the way the information about you can be synthesized and used by anyone who has the power to do so. And so the threat I think has never been greater.

And the more people spend time with AI agents and find them fun and find them usable, I think they also are going to be realizing, wow, just how vulnerable they are. You can actually say the same thing for health tech. You can say the same thing for connected cars. Day in, day out, we're getting lessons about just how much data is collected and gathered about us at any given time. And we're seeing this bear out in real time. So for the past six months, CDT's had a partnership with this nonprofit named CivAI that has this really important demo that shows what data brokers can gather and share about you at any given time. So it's a live demo. You can literally type in your own name. You can see in real time the portfolio and the dossier that could be compiled on you on commercially available data.

You also can see that an AI agent in the click of a button can compile a list of gun owners in a particular town, people that attended a particular church, people that attended a particular health clinic. We've taken that on the road to show it to lawmakers and every single person is horrified and very rightly so. And we know what the solution space looks like for that. There are very simple bills that one can pass to limit the sale and sharing of that data and put greater controls on it. And we're seeing that starting to be modeled in the states. So that's one where again, hope springs eternal, and I don't want to just be the CDT Pollyanna pushing on this, but it really does feel like there may be changing conditions on the ground for that piece of the puzzle. And so that's going to be a top priority for us.

The other piece of course is what the AI governance space is going to look like. I don't think these safety issues are going to go away. This summer showed just whether lawmakers want to or not. These crises are going to land on their desk every day with new cyber intrusions, new cyber threats, people really worrying in a very material way around what the security landscape looks like. I think the big question for a lot of us, and these are folks in the Tech Policy Press community too, is what does that package look like? What does durable effective governance actually look like? We're still figuring out what that solution space looks like and how to legislate it. And then how do we make sure that we're not just dealing with one slice of the ways in which AI impacts people's lives, but creating the very necessary space that again, I think voters are expecting and our groups and communities have been demanding of thinking about all of these places where we need better guardrails, better accountability to make sure that people's rights and interests are protected.

Brian J. Chen:

Yeah, I think there's a narrow path for some modest legislative proposals to gain traction. I mean, like Alex, I expect there to be a lot of focus on some of the frontier AI safety issues, and I would expect there to be a real dedicated push on some of those bills, some of the things that include stuff like independent AI evaluators, third party evaluators, Representative Lori Trahan's [D-Mass.] bill, the Frontier Act really is probably the one bill that's in pole position on that issue. Unfortunately, it does have issues. It has problematic language about preempting state regulation. I would imagine that there's a legislative push at the federal level on data center regulation, things like protecting utility spikes in utility rates, making sure that companies pay their own way. I mean, these things are still pretty sort of like, this stuff is still really hard. Even though that there are majorities now in the House and Senate, it is still really hard to pass a law in the US.

In the Senate, getting a majority is remarkably difficult. You have Trump who's just sort of hovering above it all saying, "If you send me something that's crazy, I'm going to veto it." And then even in the House where Dems have quite a healthy majority, in some ways that majority works against them if you care about certain very progressive proposals going forward because the majority is such that the center is going to hold. If you have a holdout on the left or on the right, they're going to have much less sway over what the body does overall. And so a situation where you have one or two senators like a Senator Manchin who single-handedly took down Build Back Better back in the day, I don't think we're going to have something like that. And anything that does move forward is going to have to be very, very carefully negotiated both within Congress as well as with the White House.

Robert Weissman:

Yeah, I guess building on that, I think there's also another dimension. So we're doing a lot of work in the states and on a variety of issues and political deepfakes, intimate deepfakes, companion chatbots, a bunch of issues. And in the states, we're seeing pretty good bipartisan support for measures. And I think that suggests, that doesn't necessarily translate federally, it hasn't, but we may be in a slightly different moment that way. That does not mean that things are going to sweep through. It just means there may be an opening and there has to be an opening to get it through the Senate unless there's some change of the filibuster, which is reasonably unlikely in the near term. So there is a bit of an opening there. That said, I think on the Democratic side, it's not just progressives versus moderates, and it may not even for this be primarily progressive versus moderates, so that's definitely a factor.

There's a decision that's going to have to be made by leadership about their willingness to offend donors, and we will see. We'll see what they're willing to do. There've been a surprising number of quotes in the media from members of Congress and even in some meetings that we've been in where members of their staff say overtly, "Yeah, but we don't want to offend the donors." And so that's still real. And although the tech side went heavily Republican, most of them actually aren't Republican or are happy to shift back. We've already seen OpenAI preemptively announce a significant pivot on how they're going to deal federally, which is to be softer and kinder in short. So we'll see. So I think in terms of specific issues, I think the AI safety frontier AI issues are definitely open because it's so pressing and it's the obviously federal issue.

I agree with Alex, it's actually complicated. It's not one where we got ready off the shelf good solutions. So that even if there's sort of political will, the policy piece is extra complicated in that space. It's possible we'll see some stuff on liability, although maybe liability just be left alone and it'll be fine. And then I do think because of some of the possibility of doing work around kids, it may be that the companion chatbot issue rises up. And then of course there's data centers, which is a little bit tricky because it's primarily a state and local issue, but there may well be some sort of federal pieces that can work on it. All that said, I think the other factor in this is I think companies, and OpenAI has already signaled this, are going to become pro-regulation. The way they were in the early days of the generative AI breakthrough moment.

They're going to come pro-regulation with super light touch regulation coupled with preemption and preempting state liability rules and regulations I think will be worth it to them to accept some really modest federal stuff. We're very worried about that deal kind of being ushered through, and I think it's certainly a thing for folks to be watching.

Justin Hendrix:

It is interesting how many folks, and more than one of you have alluded to this in this conversation, how many folks expect there to be some kind of catastrophe related to artificial intelligence in the near term? We've already seen various incidents, but perhaps nothing that quite rises to the level of catastrophe, but a lot of folks seem to suggest that is in the offing that something like that will come along. I'm interested in things like we have one questioner in the Q&A who's asked about the AI Agent Act that Senator Warner [D-Va.] put forward. Other legislation that you all are looking at or have seen that you think deserves a kind of second look in the first part of the year that would potentially address some of the concerns that could lead to a catastrophe. Anybody got anything else they're either championing or watching closely?

Alexandra Reeve Givens:

Yeah, on the question of catastrophic risks, I mean in our field it's been a highly loaded term because it has felt very much like a debate between either you are focused on catastrophic risks, which historically was extinction level, very significant mass incidents, or you're worried about current harms, which include privacy and kids and pricing and all of the things that we've been talking about. I firmly believe that we are seeing a convergence in those fields because the catastrophic risks are actually materializing not in this abstract sci-fi way. I'm not yet thinking about AI as running small city states, and I don't think that's not a good use of CDT's time. Some philosophers and professors can talk about that. But when we are seeing these cyber incidents and the increasing risk created by agents that are operating on the web, that are discovering vulnerabilities on the web, that therefore can exploit those vulnerabilities in ways that could lead to systems failure, those are now also becoming current harms.

And I don't think it is hyperbolic to say so. It's actually not buying into a hype cycle from the companies. It is actually where the evidence is going and the accountability conversation needs to go with that too. So I personally feel very much that groups like ours that have been thinking about governance and accountability and transparency and all of these elements need to be engaging on those risks as well because again, they're immediate, they are happening, and they probably are going to be the thing that most forces the political conversation in the near term if something like that materializes. It's also where you're seeing powerful actors, the deployer community, the financial services industry, critical infrastructure, all saying we need help and accountability and more resilience in that space. And I think we ignore that call at our peril because it is actually an important moment to think about the structures of responsibility and accountability that's needed for this new technological wave.

In terms of the things we're watching, it is still really important that we address those risks and we do not let that be at the exclusion of the very serious other risks to fairness, to privacy, to mental health and other places like that. And so the question is going to be, can we have legislative vehicles as well as an oversight agenda that is sufficiently capacious to address all of the ways in which AI is touching people's lives? I think we can, I think we have to, but that is going to have to be the demand of policymakers to not make it this false choice of one or the other or you only deal with this one harm. We have multiple committees with multiple jurisdiction for a reason and we have advocacy groups that can focus on different pieces of the puzzle. I do think the Frontier Act is obviously one of the places that this conversation is kind of organizing around and that's important.

It's also important to make sure that these solutions are durable and effective and have democratic accountability. I think one of the things that I've been most worried about watching things unfold over the past six months is this notion that safety and model releases just gets decided by the White House under a veil of secrecy. We have known and we have seen firsthand over the past couple of years if you needed a reminder from the lessons of history that that is not a good wielding of power and that it could easily feed into corruption and abuse and lack of public legitimacy. So we're watching his proposals and also really thinking about the democratic legitimacy and durability of those solutions. But at the same time, again, we can't ignore all of the other sector specific efforts that are going to need to happen as well, whether it's automatic decision making, the privacy pieces, fraud and abuse, and other areas like that.

Justin Hendrix:

I think I might pick up a couple of questions from the Q&A that ask you all to focus a little more in on the states. There are a couple of different questions here and some that are kind of asking what to anticipate from the new balance of power across state houses, what we can expect in terms of state AI governance and legislation. There's another question really asking about this balance of power with the federal government on preemption. Of course, the Trump administration has created this unit, the DOJ that's supposed to go out and sue the states if they pass laws that they don't like. I don't know, what do we expect just in the near term from, in some cases, newly or new trifectas across various states? Rob, maybe I'll start with you on that one.

Robert Weissman:

Yeah, I think the states have been super active and the shift towards Democrats may superpower it, supercharge it, but it's been a very lively space and I think it's responsive to what's happening and that the state legislators themselves are seeing and hearing about. So I think we're going to see accelerated action, particularly one that we're focusing on, but also where there's a ton of interest is on chatbots, the automated license readers and the Flock issue is exploding everywhere. I think we're going to see more versions of data center moratoriums or close equivalents, and I think a lot of other sort of innovations. California is really working on the AI safety issue. States may try to take down on that really is a hard one for states to have capacity to manage, and so we'll see. But I think it's going to be very aggressive in the states.

And I think that, again, I'm really worried about the federal preempt, because I don't think the Congress is going to catch up. The idea that a Congress might catch up by preempting all that the states are doing and then not do much else is very worrisome, including because I think Alex is getting exactly right in talking about both the AI safety issue itself and also its relationship to the other set of issues. I mean, we have to just have some humility about this technology that's moving super fast and thinking that we're able to, at a moment in time, come up with a regulatory structure that's going to manage it and deal with the safety issues where the technology is both complicated, the companies pushing it are as irresponsible as any company in the history of the universe. We need to be humble and not think that there's going to be a single solution around it.

And that is one reason that you just can't have the federal government preempting the states. And it's another reason we have to be really creative in thinking about what Congress does when it finally gets to this because you don't get too many bites at the apple. And once the structure is created, that's kind of what we have to live with. And particularly against the backdrop of a Supreme Court that says administrative agencies can't really evolve the authorities they have. It's going to have to be open-ended and doing a lot of learning from the last 75 years of regulatory experience and broad and humble and recognizing your humility about understanding what's going on and therefore conferring a lot of authority on a new agency to be able to do things that we can't yet anticipate are problems or necessities.

Brian J. Chen:

Yeah, and I think that the state landscape is really exciting. It's probably where you're going to see legislative proposals move forward that are unlikely to move forward at the federal level. I mean protections around worker surveillance, worker privacy, guardrails on the use of AI in the public sector, in education, consumer protection, algorithm discrimination. On the specific issue of data centers, which is so royal this electoral cycle, I mean I think this is going to be really interesting at the state level because it will be a real-time test case of that distinction between politics and policy of how far really are electeds willing to go. And I think this is where sustained organizing and sustained activism is going to be key because if that falls away, absent that, you may well have something like the lowest common denominator where lawmakers say, "We hear you and we're going to do these sort of requirements for companies to bring their own diesel generation on site, largely milquetoast things." And this really will be a test case.

The states are the laboratories of democracy. It kind of is like a laboratory for seeing how far some of this data center politics can convert into real governance.

Alexandra Reeve Givens:

The other thing I'd jump in, we've been talking about the legislative landscape, but the state AGs have been running the table recently in terms of tech governance through litigation, securing blockbuster wins that in a way creates a political precedent that you're a dummy if you're not also seeking a $25 billion award or settlement for your state through this litigation. That's being a little bit flippant, but it is a notable trend that we have really seen an uptick in recent years. And that's it's kid safety, it's privacy. Recently eight states suing OpenAI over the chatbot liability and AI liability pieces of this. And that I think is going to continue to dominate the trends, not least because some of this liability stuff is actually very difficult to legislate in writing about apportionment of liability. And so litigation is actually, I think, going to be setting a lot of the norms.

So that's another space to be watching and to something, Justin, that you and I have discussed offline, but I think it's worth bringing online, I think also raises questions about the capacity issue of just where is the technical expertise sitting right now in government? And that can be from an enforcement perspective, it can be from an expert agency perspective, how do we get enough people in CAISSI? If we create independent evaluators, where does that talent come from? How do you actually set them up for success and have them be independent? And then the question, of course, for legislators having that expertise. And frankly, as the head of a civil society organization, how do we get enough folks working in civil society that have this expertise and can go toe to toe on these issues in the way that we need to? And I do think that's an important moment, an issue for our field to be thinking about, not just in the abstract, but very acutely who should be moving into what jobs this fall to be able to be effective and help continue to navigate this seismic moment of transition for the nation and the world.

We need to be thinking about that as a field, as a pipeline, and for funders to be thinking about it as well.

Justin Hendrix:

That was actually going to be my next and almost last question. We're nearly out of time. It was really about how far off we are from being ready effectively for this kind of change and focus for these new investigations for all of this activity that we can imagine kicking off in January. But Rob or Brian, anything you want to say there on either Congress, whether it's staffed up or can get staffed up in time, what else needs to happen, what civil society needs to do to shore things up, what maybe philanthropy can do? Anybody want to speak to that?

Robert Weissman:

Well, it's obvious what philanthropy can do, be that as it may. I think the oversight thing, that's normal. Congress will get staffed up. It's always helpful to have more technical expertise, but the issues are mostly conceptual and I think the oversight's going to be really strong. I do think that there are massive challenges to enforcement and ongoing regulatory oversight if we were to ever have that in place. At the cutting edge of this, it's a very limited universe of people who understand the technology at a high level and there ain't too many of them working in government. And if you're thinking about really, if you think about the financial sector and how many bank examiners there are and how you would make the equivalent when it comes to tech and AI, leaving aside the funding for that, which Congress can do if it chooses to, identifying the people, recruiting the people or training the people, that is a massive challenge we're going to have to grapple with.

Justin Hendrix:

Brian, anything?

Brian J. Chen:

No, I mean, I think right now there is a moment I think among progressive or left of center civil society organizations to really celebrate the moment and celebrate what it means for the work that's been happening in their field and in the organizing that's been done over the last few years. And I think what's important is that this is really, it's a moment for celebration, but it is a moment also to keep up the pressure. The midterms matter in so far as it really is a moment to keep building and building to a radically different governing opportunity in 2028. And so I think keeping up that drumbeat, keeping up the organizing, keeping up a strong list of demands and not sort of falling back and negotiating away your position, I think these are things that civil society is going to have to be really strong on.

Alexandra Reeve Givens:

Brian raises a really important point there. We've obviously in this conversation been talking a lot about oversight, the accountability agenda, mitigating harms. There's also a ton of opportunity and ways in which technology can actually be useful and transformational for society and build into this vision of a government that delivers for people. And we also need talent there and we need talent that understands the lessons of history, is not just randomly techno-optimistic, but instead is deeply focused on how our social services could actually work better for people on how government use of technology could be used not for surveillance and exploitation, but to make us feel much more seamless and seen in how we integrate with government. And I share that because that movement really is important. And as we think about this question of showing that democracy can deliver for people, the Mamdani model of filling the potholes, being there at a service level, showing up for people, I don't want to completely write off tech as part of that story because I actually think it can be an important enabling part of it, but we just need to go into it with eyes open and be building that responsibly, thoughtfully with people at the center the whole time.

The same way that we need to be thinking about AI working in different languages and being culturally contextual and being relevant to people and actually being a delight. In what world can we get back to this being a delight that is enabling and helping us access more information and communicate better? And again, I'm not trying to be a Pollyanna here, but really saying that is the affirmative vision. If we want to not just be in a defensive crouch, but actually saying how is technology serving society and in furtherance of democracy, in furtherance of the goals of social justice, I think our movement needs to be grappling with that as well and having the talent and the vision to execute on that and make it an affirmative vision. So on the days when we are not just so busy doing defense, that offense and affirmative visioning has to be part of the story too, and I hope that we can make space and time for that as well.

Brian J. Chen:

Alex is so right. Harm mitigation is critical. Delivering on issues that voters care about like affordability in the immediate and near term is critical, but there is an opportunity right now to reimagine what an alternative political economy of technology could look like, what it should look like going forward. Yeah, I fully agree with Alex.

Justin Hendrix:

Well, hopefully we'll be able to figure out how to balance those things. I agree that that affirmative vision is necessary. We try to make a little room for it at Tech Policy Press, even as we often, so often are focused on some of the details of the concerns and the issues that we're looking at every day. But I'm grateful to the three of you for taking the time out this week at short notice to join us and to hopefully address many of the questions that folks had here live. This conversation will be in your podcast feeds this weekend, so look for it there. Grateful Brian, Alex, Rob, thanks so much for joining me.

Alexandra Reeve Givens:

Thanks so much.

Robert Weissman:

Thank you, Justin.

Brian J. Chen:

Thanks, Justin.