House Intelligence Committee warns of 'Black Swan' AI risks

The House Permanent Select Committee on Intelligence urged U.S. spy agencies to better prepare for "Black Swan" artificial intelligence risks, warning in a new report the technology could be leveraged by terrorists or adversaries to develop more dangerous weapons and plot more deadly attacks.

"[T]he Committee believes the [intelligence community] and policymakers more broadly must carefully consider how to prevent AI tools from being leveraged for use by rogue actors, including terrorists," the report said. It was written to assess U.S. efforts over the past quarter century to address intelligence failures that led to the Sept. 11, 2001 terror attacks.

"Preventing terrorist groups from developing access to equipment and expertise necessary to build and use weapons of mass destruction has been a significant point of emphasis for the IC and the national security establishment since 9/11," the report said. "Frontier large language models have the possibility of making it significantly easier for any rogue actor, including terrorists, to develop and conduct more destructive attacks."

The report, and warning about the destructive potential of AI, comes less than two weeks away from the 25th anniversary of the 9/11 terror attacks that killed thousands and remade U.S. security efforts. At the same time, AI is rapidly advancing and driving immense economic activity — leading lawmakers and officials to grapple with difficult questions over regulating the technology.

The report was authored by House Intelligence Chair Rick Crawford, R-Ark., and ranking member Jim Himes, D-Conn., along with Reps. Elise Stefanik, R-N.Y., and Josh Gottheimer, D-N.J. They cautioned that existing safeguards may not be enough to keep up with AI's advances.

"While AI labs seek to prevent models from returning information that could assist a bad actor in, for example, building a bioweapon, the underlying capability of models is developing rapidly and such a use of a model is difficult to fully control," the lawmakers wrote.

The lawmakers urged the intelligence community to "accelerate its own responsible adoption of advanced AI capabilities so that the United States stays ahead of its adversaries." It also said investing in "secure AI tools for collection, analysis, and warning paired with rigorous testing, human oversight, and strong privacy and civil liberties protections" is critical to addressing the next generation of threats.

The committee broadly found that the U.S. "faces a threat environment as complex and dangerous as any since September 11, 2001."

"Serious gaps remain in our ability to confront and address these vulnerabilities. We are not adequately prepared to confront and address our vulnerabilities," the report read.