OpenAI and Anthropic provided a specialized briefing to staff of the US House Homeland Security Committee on their new generation of AI models with cyber offense and defense capabilities, as well as the various impacts of such technologies on cybersecurity. This meeting represents one of the first specialized communications between US legislative bodies and the two AI giants concerning cyber threats arising from new AI models, with a particular focus on security risks facing vulnerable critical infrastructure sectors.
An aide to the committee confirmed that the two companies delivered the specialized briefings to congressional staff last Thursday in the form of two separate, non-public, classified sessions.
The aide stated that this communication was part of Congress's proactive outreach regarding the latest developments in cutting-edge AI models, primarily to assess their potential impact on the cybersecurity of critical infrastructure. The briefing content also referenced a recent White House memorandum that accused China of conducting specialized operations on an industrial scale to dismantle and replicate advanced US AI models.
OpenAI responded that, aside from the specialized meeting with the White House last week, the company had concurrently completed several similar briefings with multiple committees in both the Senate and the House of Representatives.
An Anthropic spokesperson stated that the company regularly provides updates to congressional staff on AI model capabilities and potential national security implications, characterizing last week's meeting as part of ongoing, routine communication.
Comments