The NSA is carrying out White House directives on AI, and it is holding talks with leading developers as it does so, the report said.
Speaking at an Intelligence and National Security event in Bethesda, Maryland, Kosiba said, per the report, “We want access to all the models, and we’re going to take advantage of that.”
The report said “access” could mean the NSA using a model through a company-controlled service, deploying a version inside a government environment or receiving access for security testing of classified networks.
We’d love to be your preferred source for news.
Please add us to your preferred sources list so our news, data and interviews show up in your feed. Thanks!
An executive order issued in June makes the NSA responsible for determining which AI systems are a “covered frontier model” subject to a voluntary arrangement in which developers can give the federal government access to models for 30 days before the models’ release to scrutinize potential hacking abilities, according to the report.
Another White House directive issued in June requires intelligence and defense agencies to form partnerships in the AI industry, make the most advanced models available to the agencies’ personnel and avoid becoming dependent on any single company, per the report.
During Thursday’s event, Kosiba didn’t identify the companies with which the NSA is having conversations, the models the NSA is using or whether any developers have offered the NSA access to unreleased models, according to the report.
The White House issued the executive seeking access to new AI models before their release on June 2, saying this access would allow the government to choose “trusted partners that will have early access to covered frontier models to promote secure innovation and strengthen the cybersecurity of critical infrastructure.”
About a month earlier, Google DeepMind, Microsoft and xAI each agreed to share their frontier AI models with a government agency for national security testing before offering those models to the public.
PYMNTS reported Aug. 4 that the stakes are rising because frontier models are becoming more capable of finding and exploiting software vulnerabilities, potentially expanding both defensive capabilities and the attack surface.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
Add as Preferred Source
