Democrats Have No Coherent Plan for AI

After Donald Trump's election, Democrats seized the opportunity to capitalize on the public's growing concern over AI backlash. However, now that Democrats hold a significant majority in both houses of Congress, voters expect them to follow through on their promises. Unfortunately, Democrats lack a clear plan to regulate AI. More importantly, they don't even agree on the specific problems that need to be addressed.

Federal AI legislation seemed nearly impossible until 2029, but with the recent midterms, regulation now appears more feasible. Despite some Republican lawmakers showing interest in AI regulation, they remain hesitant to defy Trump, their party leader. Of the approximately 150 AI-focused bills proposed during Trump's second term, only one passed, prohibiting nonconsensual posting of sexual images generated by AI.

With a substantial majority in the House and a potential 60-vote supermajority in the Senate, Democrats have the power to pass AI legislation. However, they still lack a coherent plan. A critical issue, data centers, has seen mixed opinions from candidates. Some argue for paying for their power bills and grid upgrades, while others suggest local approval and eliminating tax breaks.

The disagreement extends beyond data centers, as Democrats have diverse opinions on various aspects of AI regulation, ranging from job losses to deepfakes, and even the potential for superintelligent AI to cause global destruction. While some advocate for legal liability for AI companies, others suggest implementing kill switches or safety audits.

However, these solutions come with their own drawbacks, as it might be too late to prevent damage once harm occurs or knowing a model's future behavior is nearly impossible. Despite these challenges, Democrats might still be able to introduce AI regulation. They just need to figure out what specific aspects they want to target.