The TD Cowen Insight
Emerging US and EU policy for generative AI looks to protect consumers, national security, and copyright industries. For investors, Biden’s key GenAI goal is competition: scrutinizing US tech platforms’ control of key inputs (especially compute) — and perhaps embracing competition from open-source models. Therefore, the U.S. presidential election could impact tech platforms long term goals to fully commercialize generative AI.
Generative AI Regulation is Coming
Given generative AIs transformative and disruptive potential, regulation is a near given. At this early stage, we’re anticipating an overall manageable policy trajectory for leading tech companies to continue commercializing generative AI. However, competition policy in 2024 could influence long-term generative AI market structure.
United States: Aggressive Regulators But Old Laws
We suspect Congress won’t pass a comprehensive generative AI law before late 2025 at the earliest. Therefore, we expect the White House to be the main driver of generative AI policy using existing laws.
Biden’s likely 2024 actions include:
- FTC decides whether OpenAI complies with existing consumer protection law
- Commerce Department decides whether U.S. should embrace open-source foundational models
- FTC looks to limit cloud providers’ control of generative AI services
- Commerce requires future models to seek licenses before deploying
without an AI-specific statute, courts will have a big say in how far regulatory agencies can push existing laws to truly affect generative AI models/services. Keep in mind that courts are trending conservative and anti-regulation.
Impact of 2024 U.S. Presidential Election
The November 2024 election could be particularly important. A Biden win likely continues policies aimed at restraining the U.S. tech platforms from dominating generative AI. A Trump win probably lets U.S. platforms more fully leverage their considerable natural advantages (compute, data, financial resources) via M&A/partnerships — with less risk of disruptive antitrust or consumer protection rulings.
Europe: Strong New Laws But Weak Regulators
Europe appears close to enacting a new AI Act — that’s on top of three recent tech laws (GDPR, DMA, DSA). So, on paper, the E.U. is well-equipped — and it’s certainly possible they do impose aggressive generative AI measures. however, the history of E.U. tech policy (e.g., GDPR) suggests to us that Europe may have only limited effects on affecting U.S. tech platforms’ commercialization