Gautam Mukunda: Regulate AI Like the Airline Industry

Gautam Mukunda, a Bloomberg Opinion Contributor, argues that the approach to regulating Artificial Intelligence should mirror the successful model established in the aviation industry. In a discussion, Mukunda highlighted the immense economic and security implications of AI, citing its potential use in cybersecurity attacks and the development of weapons.

StartupHub data

Companies working on this

Profiles of the companies named in this story, with funding and a one-liner from our database.

Anthropic

Private / $100B+ est

Anthropic is an AI safety and research company building reliable, interpretable, and steerable AI systems, best known for the Claude family of models.

The Aviation Analogy

Mukunda pointed to the aviation industry as a prime example of how regulation can foster both safety and innovation. He recalled that 100 years ago, the idea of safe, rapid global travel would have seemed miraculous. This achievement, he contends, is inseparable from the success of regulatory bodies like the FAA, which set global safety standards. Mukunda believes this collaborative, safety-focused regulatory framework can serve as a blueprint for managing AI.

Is it Too Late to Regulate AI?

Addressing the concern that AI development might be too advanced for effective regulation, Mukunda stated that while eliminating all potential harms might be challenging, it is not too late to mitigate the most significant risks. He noted that the most severe potential harms, such as AI-powered biological weapons, are still years away, allowing time to implement regulatory barriers. He stressed the importance of establishing these safeguards before the current safety gains are lost.

The full discussion can be found on Bloomberg Podcast's YouTube channel.

China's AI Landscape and Regulation

The conversation touched upon China's advancements in AI, specifically mentioning Moonshot's Kimi K3 model. Mukunda questioned whether China would regulate AI faster than the US, noting the existence of debate within China and the country's deference to technical expertise. However, he observed that China's open-weight models, while impressive, often lag behind US counterparts and appear to be heavily influenced by American research, citing an instance where a Chinese model identified itself as Claude by Anthropic. This suggests a reliance on US-based data and development.

Mukunda also discussed China's regulatory approach, mentioning their crackdown on AI systems that discuss sensitive topics like Tiananmen Square. He characterized China's overall touch as surprisingly light, attributing this partly to their differing perspective on AI. The Chinese AI community, he explained, is more focused on practical applications like factory efficiency and short-term economic returns, rather than the existential risks that preoccupy some in the West.

Europe's Regulatory Stance

When asked about Europe's role in AI regulation, Mukunda suggested that while Europe excels at regulation, their approach is often characterized by a heavier hand. He reiterated his belief in the US model's ability to balance safety and innovation, citing agencies like the FDA as examples of successful regulatory bodies.

The Peril of Weakening Regulation

Mukunda issued a strong warning about the potential consequences of weakening government oversight of AI. He argued that companies advocating for such a stance fundamentally misunderstand the situation. If the US government becomes too weak to regulate AI effectively, it could lead to public rejection of the technology and ultimately harm the industry's growth within the country.

US Policy and Chinese Open-Weight Models

Regarding the Trump administration's stance, Mukunda mentioned discussions about banning Chinese open-weight models. He expressed reservations about this, viewing it as a potentially slippery slope. While acknowledging that these models pose a competitive threat to leading US AI labs, he also recognized their benefits for other developers and the broader AI community.