Experts are urrging governments to implement strict oversight for artificial intelligence following several incidents involving AI agents losing human control. this movement seeks to bring AI under the same regulatory scrutiny as the aviation and food industries to mitigate escalating risks.
An FDA-style watchdog for the AI era
AI regulation is moving from a theoretical debate to an urgent policy necessity. Geoffrey Hinton, a pioneer often referred to as the "godfather of AI," suggests that AI companies should be required to prove safety to an independent agecny, much like pharmaceutical companies must satisfy the U.S. Food and Drug Administration (FDA) before marketing drugs, as reported by CBC News Network.
Jaxson Khan of the University of Toronto’s Munk School of Global Affairs argues that leaving such a powerful technology in private hands without supervision is an anomaly. He notes that in no other major industry—including oil and gas, food production, or nuclear energy—do companies completely police themselves without government oversight.
The challenge of a four-month doubling cycle
The velocity of technological evolution presents a massive hurdle for lawmakers attempting to create lasting rules. One report cited in the article indicates that the capabilities of AI agents are doubling approximately every four months,a rate of change that Hinton suggests is unprecedented for any powerful technology.
This rapid advancement makes it difficult for governments to maintain relevance in the regulatory process. Manon Paquet, an associate director at McGill’s Centre for Media, Technology and Democracy, warns that by the time new legislation is put forward, the technology has often already evolved into something else entirely.
Will the US and China join a global registry?
The effectiveness of a global regulatory body remains an unverified prospect due to intense geopolitical competition. While the report mentions that 400 scientists, including Yoshua Bengio, have called for an international registry of security incidents, it remains unknown if the world's two dominant AI powers will actually comply with such a framework.
Olivier Blais of the Montreal-based consulting firm Moov AI points out that without the active participation of the United States and China, any international oversight body may lack the authority to be truly useful. This tension is exacerbated by U.S. President Donald Trump’s dismissal of AI warnings, as he prioritizes staying ahead of China in the ongoing AI arms race, according to the report.
A 400-scientist push for international containment
A coalition of leading scientists is attempting to build a safety net through international cooperation.. This group, which includes both Hinton and Bengio, is advocating for an international body to oversee containment standards and the reporting of incidents involving advanced, cyber-capable AI systems.
The proposed framework would include a public registry of significant security incidents to ensure transparency. jaxson Khan emphasizes that while a perfect model of global governance may be impossible to achieve , establishing a stable, shared system is essential to prevent the potential misuse of these increasingly capable systems.
Comments 0