Abstract:
On September 1, according to CNBC, the U.S. House of Representatives Permanent Special Intelligence Committee urged U.S. intelligence agencies to better guard against “black swan” AI risks. The committee warned in a new report that terrorists or hostile forces could use the technology to develop more dangerous weapons and plan more deadly attacks.
One World Trade Center is located on the site of the 9.11 attacks
"The committee believes that the intelligence community and broader policymakers must carefully consider how to prevent AI tools from being exploited by malicious actors, including terrorists," the report said. The report aims to assess the steps the United States has taken over the past 25 years to respond to intelligence failures that ultimately led to the terrorist attacks of September 11, 2001.
"Since 9/11, preventing terrorist groups from acquiring the equipment and expertise needed to build and use weapons of mass destruction has been an important priority for the intelligence community and national security agencies. Cutting-edge large-scale language models have the potential to make it easier for any malicious actor, including terrorists, to plan and carry out more devastating attacks," the report said.
The report and its warnings about AI’s destructive potential come less than two weeks after the 25th anniversary of the 9/11 terrorist attacks that killed thousands and reshaped the U.S. security system. At the same time, AI is developing rapidly and driving huge economic activity, leaving lawmakers and officials to grapple with thorny issues surrounding the regulation of the technology.
The report was co-authored by House Intelligence Committee Chairman Rick Crawford (R-Ark.) and committee ranking members Jim Himes (D-Conn.), as well as Reps. Elise Stefanik (R-N.Y.) and Josh Gottheimer (D-N.J.). They warn that existing security safeguards may not be enough to keep up with the rapid pace of AI development.
"Although AI labs attempt to prevent models from providing information that could help criminals carry out related actions, such as creating biological weapons, the underlying capabilities of the models are rapidly evolving, and it is very difficult to fully control such uses of the models," the lawmakers wrote.
These lawmakers urged the U.S. intelligence community to "accelerate the responsible adoption of advanced AI capabilities to ensure that the United States remains ahead of its adversaries." The report also notes that investing in “secure AI tools for information collection, analysis, and warning, paired with rigorous testing, human oversight, and strong privacy and civil liberties protections,” is critical to combating the next generation of threats.
The committee generally believes that the United States "faces the most complex and dangerous security threat environment since September 11, 2001." "We continue to have serious gaps in our ability to respond to and address these vulnerabilities. We are not fully prepared to address and address our vulnerabilities," the report said.
Comments