The cyber capabilities of Anthropic’s Mythos and OpenAI’s GPT-5.5 show that there is a need for the government to develop a policy response that addresses the risks around frontier artificial intelligence models while promoting secure innovation, according to the Institute for AI Policy and Strategy (IAPS).
“Risks beyond cyber, including biosecurity threats, are likely to emerge without additional safeguards,” IAPS researchers Joe O’Brien, Brianna Rosen and Christopher Covino wrote in a May 14 IAPS Policy Memo. “A strategic policy response must look beyond the Mythos moment to secure models against adversaries, drive defense through automation, expand public-private information sharing, and build government capacity.”
It was reported April 7 that Anthropic was allowing select partners to gain early access to its Claude Mythos Preview so that they could identify vulnerabilities and strengthen systems before threats could be exploited. On Friday (May 22), Anthropic said its Mythos Preview model had ferreted out more than 10,000 cybersecurity vulnerabilities.
OpenAI said April 14 that ahead of its release of increasingly more capable models, it was expanding access to a program that gives cybersecurity professionals access to frontier models.
In the IAPS memo, the IAPS researchers said that to secure strategic AI assets against theft and sabotage by adversaries, federal agencies should accelerate the development of technical standards for high-security data centers, coordinate with industry to test advanced security measures, and extend partnership models to include infrastructure housing frontier model weights for voluntary threat detection and analysis. The agencies should also expand their intelligence insights into adversarial development and use of frontier AI, the memo said.
IAPS also suggested in the memo that to defend against AI-enabled cyber operations, federal agencies should scale defensive automation and build infrastructure to detect and disrupt offensive cyber agents. They should also automate defensive research and development (R&D) across high-priority safety and security domains, per the memo.
Advertisement: Scroll to Continue
To track frontier AI risks, agencies should create a central federal hub for AI risk information sharing; expand public-private information-sharing mechanisms; provide agency-specific guidance on agent identifiers to track and monitor federal AI use; and invest in evaluation science and a third-party evaluation ecosystem, the memo said.
IAPS also recommended in the memo that the White House strengthen the federal government’s ability to prevent and respond to AI-enabled crises, and that it empower the National Institute of Standards and Technology’s Center for AI Standards and Innovation (CAISI) to meet the national security demands of frontier AI.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
IAPS Recommends Federal Action to Secure Frontier AI Models From Theft | PYMNTS.com Top World News Today.
Hence then, the article about iaps recommends federal action to secure frontier ai models from theft pymnts com was published today ( ) and is available on TOP world News today ( Middle East ) The editorial team at PressBee has edited and verified it, and it may have been modified, fully republished, or quoted. You can read and follow the updates of this news or article from its original source.
Read More Details
Finally We wish PressBee provided you with enough information of ( IAPS Recommends Federal Action to Secure Frontier AI Models From Theft .. PYMNTS.com )
Also on site :