Opinion
It's Time to Regulate Frontier AI the Way We Regulate Drugs.
—
By Michael Posner
As seen in: Forbes
The need for government regulation of Big Tech became much more obvious earlier this month when Anthropic announced the development of a new artificial intelligence model called Mythos, a product capable of finding and exploiting hidden flaws in the software that runs our banks, power grids and governments. Anthropic delayed its public release, instead sharing Mythos with more than 40 organizations, including companies like Apple, Microsoft and Amazon, that provide technology used in maintaining critical global infrastructure. It urged each of them to develop security fixes for the vulnerabilities Mythos has identified. It is unclear whether and, if so, when Anthropic will release Mythos to the broader public. In the U.S., at least, there is no federal regulatory agency with the authority to set guardrails relating to the introduction of new technologies with dangerous side effects.
The Trump administration has shown no interest in filling that regulatory gap. To the contrary, President Trump has pursued a policy of deregulating frontier AI development, and his allies in Silicon Valley—among them Elon Musk and Peter Thiel—have actively worked to dismantle what little federal oversight exists. The result is that the United States federal government is not only failing to regulate Big Tech but is moving in the opposite direction.
Other democracies have not stood still. The European Union has taken the lead in attempting to govern AI and the digital economy, most notably through its AI Act, which came into force in 2024, and its Digital Markets Act. As Mariana Olaizola Rosenblat has written, the EU framework is imperfect and faces real implementation challenges, but it represents something the U.S. currently lacks: a serious attempt to balance innovation against public safety and democratic accountability. While California and New York have enacted laws that require the disclosure of frontier model capabilities, they do not give regulators the authority to pause the deployment of a model or require mitigations. Even so, the Trump administration has been trying—unsuccessfully to date—to preempt those states’ attempts at regulation.
Read the full Forbes article.
___
Michael Posner is the Jerome Kohlberg Professor of Ethics and Finance, Professor of Business and Society and Director of the NYU Stern Center for Business and Human Rights.