The Midterms and AI Development

The midterm elections have resulted in a crushing defeat for the Republican Party and one of the most impressive and resounding Democratic party victories in recent memory. I suspect that both parties will spend the coming weeks analyzing and processing the results and what they mean for American politics. The consequences of this election will likely change the political landscape in fundamental ways both in the near-term (over the next two years), and in the lead up to the 2028 election. I wanted to spend some time analyzing the political ramifications of this shift, especially in as far as it affects the politics of AI development.

If you disagree with any of my statements please feel free to leave a comment. I am by no means a political analyst and may have made mistakes. All criticism is welcome!

Why this matters?

Many of those closely following AI development believe that we are rapidly approaching the critical milestone of recursive self-improvement (when AIs can fully automate most or all of AI R&D potentially triggering a rapid self reinforcing explosion in AI capabilities). If these forecasts are correct, then the exact political situation in the United States (where much of the world's frontier AI research is concentrated) becomes of paramount importance.

What Happened?

The midterm elections in the United States take place two years into a president's 4 year term with all 435 members of the House of Representatives up for election (218 needed for a majority) along with a third of the Senate (35 out of 100 Senators stand for election, 51 needed for a majority). In addition, the midterms involve voting for State level and Local officials (governors, state legislative bodies, etc.). Generally, the midterms are viewed as a referendum on how voters feel about the current ruling party/administration (most voters cast their ballot for the party they like rather than evaluating candidates individually), and usually result in the ruling party losing seats.

Yesterday, the midterm elections for the second Trump administration took place, and though votes are still being counted in some races, it is already clear at this point that voters have shifted decisively towards the Democrats (Blue Wave). Democrats are currently projected to win ~55 seats in the Senate (4 more than is needed for an outright majority) and a stunning ~258 seats in the House (40 more seats than is needed for an outright majority). If these projections hold, this result would be on par with the 2006 Democratic midterm victory, when voters handed Democrats a sweeping victory to express their dissatisfaction with the Iraq war and the Bush administration.

Who are the winners and losers?

The most obvious winners are the Democrats across the board, but especially the progressive wing of the Democratic Party. It is important, however, not to overstate the progressive victory. Although progressive candidates have achieved impressive victories in purple states and expanded their representation in both chambers of Congress, mainstream Democrats have also won decisively and still maintain near total control over the party apparatus. In fact, the scale of the Democratic victory in both chambers will likely increase the power of the mainstream wing of the party relative to the progressives, since the latter will have less leverage to block legislation or extract concessions from party leadership.

On the other side of the aisle, there are very few winners. Republicans have been routed across the board. This election heralds the beginning of the end of the Trump era, forcing Republicans to confront the question of what their party will look like in a post Trump world. In the coming weeks, the blame game will begin and different factions in the Republican party will try to pin the blame on each other. Tentatively, and based on initial reactions, it appears that this election may be a relative victory for the populist/nationalist wing of the party represented by Tucker Carlson and others in his orbit who have broken with Trump openly and may now seem vindicated in their criticism of his administration. Still, this is by no means guaranteed, and it will be important to watch the jockeying for power and narrative control within the Republican party in the coming weeks to understand how the balance of power is shifting.

AI Skepticism in Campaigns

The midterms have forced both parties to contend directly with their positions on AI development. Though neither party has yet developed a very well defined or coherent position on this issue, Democratic candidates have tended to express a more AI skeptical message. In many races, Democratic candidates (especially more progressive Democrats) have woven datacenter and job worries into their messaging. Republican candidates on the other hand have been more confused on this issue with more populist Republicans presenting an AI skeptical message while mainstream Republicans uphold the Trump administration's more pro-datacenter/pro-business line.

Though much rarer, some races have also featured candidates speaking more openly about AI existential risks, bringing X-risk discussions for the first time into mainstream electoral politics.

It is unclear to what degree AI Skepticism contributed to Democratic victory. Although polling does show that the electorate broadly favors the Democratic Party's policies and rhetoric on AI, polls also show that AI policy is not necessarily in the list of top priorities for voters especially compared to other bread and butter political issues like the economy, immigration, and the Iran war.

Political Repercussions for the AI Industry in the Next Two Years

Overall, the midterm election results are likely to lead to a noticeable shift toward a more pro regulation stance on AI. The Democrats, being the party of opposition, and having directly campaigned against AI datacenters and job loss, may be increasingly inclined to embrace AI skepticism. Democrats in Congress are less likely to view themselves as responsible for the US economy (US voters generally hold the president's party accountable for the economy) and so may be less susceptible to arguments that AI regulation will harm the economy. Instead, Democrats will likely position themselves as the pro AI regulation party and will cast any labor disruptions or AI warning shots/catastrophes in the next two years as a direct result of the administration's reckless policies.

The Balance of Power Between Anthropic and the Administration has shifted

If there is one Silicon Valley company that benefits from these election results, that company is Anthropic. Even before the election, Anthropic was emerging as the dominant company in the AI race. They have captured a large and growing segment of the AI industry, they have produced the most powerful and commercially safe (at least relative to their main competitor OpenAI) models, and they are within sight of a massive 2 trillion dollar IPO which could provide the company with an enormous infusion of capital, allowing it to further outpace its competitors.

The main serious threat to Anthropic came from the Trump administration which clearly viewed them with suspicion and which could potentially have used legal and regulatory obstacles to harm them (they had already done this to some degree). Now, with Congress firmly under Democratic control, the administration's ability to exert pressure on Anthropic may be significantly diminished. Moreover, as the administration faces intensified scrutiny and opposition from the Democrats in Congress, it may not have the time, political capital, or appetite to confront Anthropic directly.

Democrats vs Elon Musk

On the other side of the coin, this election is a significant blow to Elon Musk who has closely aligned himself with the Trump administration. Elon spent massively to support the Republicans in the midterms. With Democrats now firmly in control of Congress, they face an important strategic decision on how confrontational to be towards Musk. Do they directly pursue him aggressively with their oversight powers which risks hardening him into an inveterate political enemy, or do they adopt a more conciliatory approach in hopes of improving relations with him? It seems likely, however, that many Democrats in Congress will pressure their party leadership to take a more aggressive stance towards Musk, pushing to subpoena him and investigate his political activities, businesses and his involvement with DOGE.

The Chances of Passing Comprehensive AI Regulation Legislation: Very Low -> Low

It is important to note that, barring a change in stance by the Trump administration, this result far from guarantees any substantive legislation on AI development in the next two years. Although Democrats, with strong majorities in both chambers of Congress, may be able to push through some form of legislation (even this is uncertain, given that Republicans can still utilize the filibuster in the Senate), the Trump administration still retains the power to veto legislation. To override the presidential veto, Democrats will need a two thirds majority in both chambers of Congress which would require them to secure roughly 32 defections amongst Republicans in the House, and 12 defections amongst Republicans in the Senate. It is therefore unlikely that congressional Democrats will be able to push through any comprehensive AI regulation legislation unless
1. The Trump administration changes its stance on AI regulation
2. The Republican Party is beset by so much infighting and factionalism that a substantial fraction of congressional Republicans are willing to break with the administration on this policy.

Limited AI Regulation Legislation: Low -> Moderate

Although the chances of passing comprehensive and restrictive AI legislation remain low, prospects for passing more targeted and less contentious measures have improved considerably. Such legislation could include proposals like the Kill Switch Act, regulations governing AI datacenter construction, or more narrowly targeted legislation to address AI related cybersecurity and bio security risks. These measures are much more likely to pass if they are incorporated into broader bipartisan legislative packages and exchanged for some Republican legislative priorities. Such compromises may be particularly feasible since some Republican congressional representatives have themselves expressed AI skeptical views, while others may not regard these issues as particularly consequential to their constituents.

Congressional Investigations, Subpoenas, and Public Messaging: High

Perhaps the most significant expected change resulting from the election will be Congress's increased willingness to exercise its oversight powers to launch investigations, hold hearings, and compel the release of documents related to AI development. Democrats may have strong political incentives to pursue such investigations especially to target companies and individuals closely associated with the Trump administration including David Sacks, Elon Musk, Sam Altman, and Jensen Huang. The political goal of such hearings and investigations would likely be to represent the Trump administration as corrupt, captured by powerful billionaires, and incapable of responsibly managing the risks associated with AI development.

Funds for AI Safety in the Federal Budget: Moderate

With their power over the purse, Democrats could exercise considerable influence over the federal budget. They could seek to allocate funding to AI safety orgs and third party auditors. Such funding could be incorporated into broader spending bills as part of budget negotiations with Republicans.

The Chances of an AI Treaty: Very Low

Treaties with foreign powers must be ratified by the President. Therefore the recent democratic victories in Congress would not directly increase the chances of signing an internationally binding AI treaty.

Gridlock: High

Unfortunately, the most likely political outcome of this election is a deeply divided and gridlocked federal government. It is therefore unlikely, all else being equal, that the government will be able to play a coherent and active role in managing AI development over the next two years. Although we can plausibly expect greater transparency into AI development due to congressional oversight and if we are lucky limited legislative successes in implementing non-restrictive AI legislation or greater federal funding for AI Safety, the government will likely be too divided to do much unless a crisis compels the executive to take unilateral action.

Repercussions on the 2028 Election

The effect of this midterm election will almost certainly carry over into the 2028 presidential election. However, precisely how these midterm results will shape that election is uncertain and will depend on how the political landscape evolves over the next two years. Nevertheless, there are several important trends worth monitoring that could provide early indications of how the winds are blowing.

Mainstream vs Progressive Democrats

The first and perhaps most important trend to watch is the balance of power between the mainstream and progressive wings of the Democratic Party. Progressive Democrats are much more vocal in their skepticism toward AI development. If populist, progressive Democrats are able to strengthen their position in the party and exert greater influence over Democratic policy, this could set the stage for a progressive Democratic presidential nominee in 2028. Such a candidate might campaign on a much more AI skeptical platform, and, if elected (which would seem likely given historical precedent), would be empowered to pursue much more aggressive policies in 2029-2032.

The Populist Right Wing and Their Views on AI Development

Another factor to track is the strength and the policy positions of the populist right wing in the Republican party. As mentioned earlier, though the results of this election on the right wing are still uncertain, it seems plausible that this result will facilitate a shift in power towards the more populist right wing. Unlike the Republican establishment, the populist right has voiced more AI skeptical views. It would be important to track this faction's strength in the Republican party and their positions on AI development. If this wing of the party grows in power and hardens towards more AI skeptical positions they may compel Republican nominees for 2028 to campaign on more AI skeptical positions. If both the Democratic and the Republican nominees are AI skeptical, the presidential election campaign could trigger a political feedback loop in which both parties compete to demonstrate their willingness to regulate AI development, pushing the broader political debate toward increasingly interventionist positions.

The Administration and JD Vance's Political Positioning

Finally, another important development to watch is whether the Trump administration shifts its stance at all on AI development, and particularly whether JD Vance begins articulating more AI skeptical views (whether or not these attitudes are reflected by others in the administration). As heir apparent of the Trump movement, JD Vance's positions on AI development could shape the Republican Party's future stance on the issue or serve as a foil for rival Republican candidates seeking to distinguish themselves from him.