From the Hilltop to Capitol Hill: Framing the Future of Open vs. Closed AI Policy at Georgetown

Posted in Events News  |  Tagged , ,

By Alex Dragone MBA/MSFS’27

As Congress determines the fate of artificial intelligence regulation, the Georgetown Center for Business and Public Policy hosted a roundtable of artificial intelligence experts in the Rayburn House Office Building on June 9. With the American AI industry seeking federal support, and Chinese competitors demonstrating their capabilities, Congress can push American AI to develop proprietary products, only modifiable by their makers – “closed” models; or those which users can manipulate and “open” models.

“The cat is out of the bag” on open-source modeling, noted Frank Nagle, strategy professor at Harvard Business School and advising chief economist to the Linux Foundation. Any attempt by Washington to restrict open-source development would bring about a massive competitive disadvantage for the U.S. relative to China. DeepSeek, a Chinese AI start-up, employs open-source software, and was able to train its large language model at a far lower cost and with far less chips than American firms like OpenAI, Google, and Meta had done. Karl Zhao, a generative AI consultant for DeepSeek, pointed to OpenAI CEO Sam Altman’s comment in 2023 that it would be “totally hopeless” for a firm with ten million dollars to compete with OpenAI. DeepSeek has shattered that claim.

While DeepSeek’s advances seem to offer a way for firms based in less prosperous countries to compete in the AI space, Nagle was skeptical, as these firms are still reliant on an ecosystem of physical and social infrastructure, e.g., servers and universities. Nagle warned that AI, rather than uplifting the developing world, could lead to greater global inequality.

The subject of an “arms race” between the U.S. and China came up repeatedly during the panel discussion. Any considerations in Washington or Beijing to regulate AI could evaporate when faced with the prospect of falling behind the other in technology and the resulting economic gains. Congress is already feeling this pressure; the Senate is now considering the House of Representatives’ version of the “Big, Beautiful Bill,” which bans states and their subordinate polities from regulating AI for a decade. Jessie Wang, an economist at RAND Corporation, said one way to avoid a race to the regulatory bottom is to institute a global framework to control AI. Mike Linksvayer, head of developer policy at GitHub, offered that the most important thing the U.S. could do to regulate AI would be to adopt good general governance policies, making federal bodies quicker to respond to rapidly changing developments.

Regarding principles that should shape AI development, Wang advocated a balance between the poles of total control and absolute openness. Zhao suggested Isaac Asimov’s three laws of robotics.

The tensions in the AI debate – between open and closed models, between national competition and international cooperation, and between regulation and laissez-faire – will shape debates in the halls of power, the executive suite, and laboratories for years to come.

Watch the roundtable highlights: