Cozen Currents: Okay Doomer

“While Washington, DC, as well as the states, have become fixated on the potential existential threat posed by AI, no consensus exists over how – or even whether – to regulate it.” — Howard Schweitzer, CEO, Cozen O’Connor Public Strategies

The Cozen Lens

• Doomerism is dominating the debate over regulating AI with different implications for major labs and smaller firms.
• While political support for data centers continues to collapse, politicians are trying to walk a line between appearing tough on AI and not scaring away investment.
• When President Trump and Chinese President Xi Jinping meet in Washington this week, the summit will prioritize stability and tangible deliverables will likely be relatively modest.

Subscribe to Cozen Currents

The Politics and Policy of AI Doomerism

AI Safety is Having a Moment. Concerns about existential AI dangers are going mainstream.

• Worry about AI’s existential risk to humanity, once the province of Silicon Valley “doomers,” has spread to Washington with lawmakers expressing worries about AI safety this month. It cuts across the red-blue divide: both Senator Bernie Sanders (I-VT) and top Trump ally Steve Bannon evoked similar themes in speeches at an AI safety conference last week.
• Anthropic CEO Dario Amodei, OpenAI CEO Sam Altman, and xAI CEO Elon Musk all agreed on the need to pace AI development, notable given that these three AI leaders seem to have no love lost for each other. While Anthropic has long advocated for AI rules, OpenAI recently shifted its position. The company now backs a requirement for independent AI safety monitors, a feature of Reps. Jay Obernolte (R-CA) and Lori Trahan’s (D-MA) FRONTIER Act. This shifts from its stance of “reverse federalism,” which entails supporting the passage of similar laws in different states to create a de facto national standard. The Information reported that Anthropic, OpenAI, and Google have discussed creating a standards-setting body for AI.

The AI Sector isn’t a Monolith. The implications of regulating AI safety aren’t the same for different labs.

• The AI safety debate is unfolding amid competition among major labs offering proprietary, closed models and smaller firms. Some labs offer open models, including both Chinese and American firms. Slowing AI development and tightening government regulation could affect leading AI firms and newcomers differently.
• Bigger labs with more resources could navigate new regulations more easily, while compliance could be a greater burden for startups. Tighter rules for AI safety, depending on how they’re written, could disproportionately affect smaller firms.
• Some opponents of AI guardrails critique the situation as an attempt at regulatory capture.

What’s Next? The AI safety movement clashes with President Trump’s laissez-faire approach to AI.

• The Wall Street Journal reported Chief of Staff Susie Wiles, Treasury Secretary Scott Bessent, and other officials have been meeting regularly to discuss AI guardrails. For now, Trump remains behind a light touch approach. In social media posts last week, he rejected the idea of tightening AI rules and dismissed claims that AI poses an existential risk as a “hoax.”
• House Democrats are considering focusing on AI in the next Congress if they retake control of the chamber. Politico reported that they are considering a select committee on AI with subpoena power that could shine a spotlight on the issue. An AI select committee could call in tech CEOs for hearings and serve as a clearinghouse for legislation.
• While AI risk may be more abstract than visceral for voters right now, it may become a bigger political issue in 2028 as AI further evolves. Some potential candidates for the 2028 Democratic presidential nomination have begun staking out positions on AI. Governor Josh Shapiro (D-PA), a 2028 contender up for re-election this November, gave a speech last week calling for federal requirements for third-party oversight and transparency of AI models. Governor Gavin Newsom (D-CA), who has at times acted as an ally of Silicon Valley, has shifted his stance by signing bills to enact guardrails this month. Alex Bores (D), the New York state legislator who sponsored the Empire State’s frontier AI legislation and ran unsuccessfully for Congress this year, is seeking to influence the debate over AI safety with a new political advocacy group.

The Politics and Policy of Data Centers

The Politics Are Toxic. Data centers have become a flashpoint in the backlash against AI as they are the clearest physical embodiment of AI. Nevertheless, the tsunami of development plows on.

• Support for data centers continues to collapse. Seventy-five percent of Americans now oppose building data centers in their communities, according to an August Heatmap Pro poll. By contrast, just a year ago, that (exactly worded) question was answered with 43 percent of the population supporting such development and only 42 percent opposing it. The data center backlash is now mainstream.
• Politicians, once behind the curve, are indeed running as fast as they can away from the falling knife. Candidates from both parties are attacking opponents for being too friendly with the industry, or not tough enough on it. New York has a moratorium on new large centers, while Texas and Massachusetts have enacted similar pauses. Other states may follow suit. The cancellation rate for disputed data center projects has risen to nearly 50 percent. Only half of the AI computing capacity planned between now and 2028 is on schedule. Local pushback is canceling billions of dollars of projects, and The Information records more than 500 pauses or bans around the country as of early August (more than 150 were added in July).
• But while the marginal data center may be getting canceled and the cost of capital to build the average center may be rising, the tsunami of investment continues. Spending on related construction increased to over $75 billion in June, an almost 60 percent year-over-year increase. As the Federal Reserve noted last month, the AI buildout largely concerns “price-insensitive technology companies [which] are more concerned about the ‘speed to power’ than the price of power.”

Finding Policies to Assuage Everybody. The federal, state, and local levels are all approaching their own equilibrium points in their responses to data centers.

• Far from being a source of concern, the federal government is encouraging AI center developers, with President Trump among the biggest AI boosters. The Trump administration is making it easier for them to build and is offering up public land to do so, where state and local opposition has little sway. President Trump’s principal response is a non-binding commitment from the private sector: the Ratepayer Protection Pledge, which the biggest AI firms signed in March and to which a few hundred more utilities and developers have since added their names. The Trump administration will regulate data centers (and AI in general) as little as it can, given the political environment.
• Congress is barely a concern because it is at an impasse. Both parties are still wrapping their arms around the AI issue, with pro-development figures and populist skeptics on each side of the aisle. With Trump still pulling his party towards deregulation, no binding limits have a chance of getting over the Senate filibuster or the president’s veto, even if Democrats flip both chambers in November. The marquee legislative initiative is the Ratepayer Protection Act, which would codify a part of the voluntary pledge into law. It passed the House 417-3 last week but required unanimous consent to proceed in the Senate. There, Senate Energy and Natural Resources Committee Ranking Member Martin Heinrich (D-NM) blocked it. Heinrich cited the fact that the bill does little more than kick the issue right back to the states, arguing it didn’t go nearly far enough. Democrats who agree with the policy in principle may also be reluctant to give Trump and Republicans an easy way out on the data center issue, given how little impact the policy may actually have.
• At the state level, the broad rules of the road are firming up. Democratic governors in Pennsylvania, Michigan, and New Jersey are circulating lists of requirements for prospective centers to follow. Some of the items are specific to the left side of the aisle, such as the use of union labor in construction or renewable energy guarantees. Republican states have been a bit less gung-ho but are still adopting restrictions. GOP governors have widely signed onto the principles of the nonbinding Ratepayer Protection Pledge. On both sides of the aisle, measures under discussion include disclosure requirements for water and energy use, as well as requirements for community benefit agreements or input.
• States could impose further data center moratoria, but if they do, they’re more likely to be in states where the move is mostly symbolic. Maine, for example only had a single data center under development when their state’s bill was vetoed. The real battleground in the data center fight will be the restrictions they face in building and the eligibility requirements for state tax incentives. States are still competing against each other to attract AI investment and tax policy provides states with the most effective tool for influencing the sector. At least eight states restricted or repealed their tax breaks this year.

What to Expect from the Trump-Xi Summit

The Stability Summit. Chinese President Xi Jinping’s return to Washington after more than a decade is an opportunity to build on the stability both sides have sought to maintain since May, even as the relationship continues to test its own red lines.

• When President Trump and Chinese President Xi Jinping meet later this week in Washington, it will be Xi’s first visit to the US since 2017 and his first state visit to Washington since 2015. It may also not be the last time they meet in person this year. China is scheduled to host the Asia-Pacific Economic Cooperation summit in November and the US is slated to host the G20 leaders’ summit in December. Both events could offer opportunities for further in-person meetings between the two leaders.
• The summit’s overriding theme will likely be stability, building on the tone set when Trump and Xi met last May in China. While the relationship has experienced periods of tension over the last several months, the growing mutual understanding of each other’s red lines has enabled Trump and Xi to prevent the relationship from deteriorating.

Trade Truce Renewal Tops the Agenda. Trump and Xi are expected to renew their trade truce and add specifics on a Board of Trade aimed at clearing tariffs on $30 billion in non-sensitive goods on each side.

• At the top of the to-do list when Trump and Xi meet will be renewing the trade truce the leaders agreed to last fall. There is little doubt the two leaders will extend the deal that limits tariffs and export controls between the two countries, but questions remain about how long the extension may last, with some suggesting it could run through the end of Trump’s term.
• In addition, Trump and Xi are expected to announce more details around the Board of Trade the leaders agreed to pursue last spring. The body aims to reduce tariffs on $30 billion worth of non-sensitive exports from each country. Both the US and China have taken steps toward implementing these changes since Trump and Xi’s last meeting, but they have not released details on exactly what may be included. Some reports have indicated that covered goods could include American energy and agricultural products.

Why Neither Side Will Blink on AI. Despite AI’s rise as a flashpoint issue, the summit is unlikely to produce binding commitments, instead repeating prior rhetoric on responsible use and further dialogue rather than real constraints.

• While AI has become an increasingly hot-button political issue, the Trump-Xi summit is unlikely to produce significant new policy changes or statements from either country. Rather than reaching any specific agreements on AI, the two countries are more likely to reiterate points from the last summit, encourage the responsible use of the technology, express a desire to create shared safety standards, and pledge to pursue further bilateral AI dialogues.
• Part of the challenge in getting Trump and Xi to make the sweeping commitments on limits to AI use that AI safety advocates have called for is that both countries see each other as their greatest rival in driving AI development. Trump specifically highlighted competition with China in pushing back against recent calls to slow AI development. Meanwhile, Chinese officials argued that these calls were motivated as much by a desire to contain Chinese competitors as by genuine safety concerns.

Share

Share this...

Facebook

Pinterest

Twitter

Linkedin