The Davis Institute for Artificial Intelligence hosted Paulo Carvão last Thursday (Feb. 19) to analyze the current landscape of artificial intelligence regulation. Carvão focused his lecture on the divergence between American and European legal frameworks while identifying the economic and social risks associated with rapid technological deployment. Throughout the talk, he explained how governments attempt to balance market innovation with public safety.
Carvão began by providing historical context for American regulation, noting that the United States has relied on a hybrid system for nearly 150 years. He said that this model allows the market to create solutions by default while using regulation to correct harms after they occur. To illustrate this approach, Carvão identified antitrust laws and litigation as the primary tools for intervention. He contrasted these mechanisms with sectoral regulation, which establishes specific industry rules before problems arise.
The presentation further highlighted a growing gap between the United States and the European Union. Carvão explained that Europe pursues structural rules through legislation like the EU AI Act to set standards for digital platforms and automated systems upfront. In contrast, the United States federal government remains divided. Because congressional dysfunction has forced a reliance on executive actions, Carvão stated that policy stability is threatened by the potential for major shifts every four years.
Apart from legal structures, Carvão identified three major limits to the current trajectory of AI development. The first limit is financial, as the industry currently spends $100 billion on infrastructure while only generating ten billion dollars in revenue. According to Carvão, this gap suggests that current investment levels are unsustainable. The second limit involves physical constraints, such as the vast amounts of land and electricity required for data centers. These needs have also already triggered environmental pushback from local communities, he said.
The final limit is moral and political. Carvão argued that AI has become a kitchen table issue affecting jobs and utility costs for average families, which will likely make it a central topic in the 2026 elections. He noted that lobbying from tech companies has increased significantly, creating a risk that the industry could become politically untouchable like the oil or pharmaceutical sectors.
During the question and answer session, students asked about practical steps individuals can take to influence these trends. When one student asked how a single person can affect a massive global industry, Carvão replied that voting remains the most effective lever. He urged attendees to research candidate positions at both federal and representative levels while also suggesting that consumers vote with their dollars through their subscription choices.
Another attendee asked about the risk of social backlash if technology develops faster than the law can manage. Carvão acknowledged that civil unrest is a legitimate policy risk and suggested that slowing the pace of development could reduce social friction. He argued for a dynamic governance model that uses iterative cycles to set narrow goals and measure compliance through audits and sandboxes.
Carvão concluded by expressing a balanced view of the future. While he remains skeptical of the current investment bubble, he expressed hope that electoral pressure could produce meaningful compromises. He emphasized that the next few years will determine whether AI serves the public interest or operates without oversight.
~ Stephen Owusu Badu `27
Be First to Comment