Microsoft, Google, xAI will give US access to AI models for security testing | Business and Financial Issues

The deal comes days after the Pentagon announced a deal with seven tech giants to use AI in select systems.

Tech giants Microsoft, Google and xAI have said they will allow the United States government access to their new types of artificial intelligence for national security experiments.

The Commerce Department’s Center for AI Standards and Innovation (CAISI) announced the deal on Tuesday amid growing concerns about the potential that the newly unveiled Anthropic Mythos version could offer to hackers.

Under the new agreement, the US government will be allowed to inspect samples before they are shipped and conduct research to assess their potential and safety risks.

The agreement fulfills a promise made by US President Donald Trump in July to cooperate with technology companies to assess their AI models for “national security risks”.

Microsoft will work with US government scientists to test AI systems “in ways that explore unexpected behavior”, the company said. Together they will create a shared data and testing platform for the company’s brands, the company said.

Microsoft has signed a similar agreement with the United Kingdom’s AI Security Institute, according to a statement.

Concern is growing in Washington over the national security risks posed by powerful AI systems. By providing access to border personnel, US officials want to identify threats ranging from cyberattacks to military misuse before weapons are deployed in large numbers.

The development of advanced AI systems, including Anthropic’s Mythos, in recent weeks has caused a stir around the world, including among US officials and American companies, because of their potential to increase the number of hackers.

“Independent, rigorous measurement science is essential to understanding the limits of AI and its implications for national security,” CISI Director Chris Fall said in a statement.

The move is based on the 2024 agreements with OpenAI and Anthropic under the leadership of President Joe Biden while CAISI is known as the US Artificial Intelligence Safety Institute. Under Biden, the agency focused on developing AI tests, definitions and voluntary defense standards. It is led by Biden’s technology adviser Elizabeth Kelly, who joined Anthropic, according to her LinkedIn profile.

CAISI, the government’s main body for AI prototype testing, said it has already completed more than 40 tests, including advanced models not available to the public.

Developers often offer their models with security features that can be removed so that the site can investigate national security threats, the agency said.

xAI did not immediately respond to a request for comment. Google declined to comment.

On Wall Street, Microsoft shares fell 0.6 percent in afternoon trading after the announcement. Alphabet, Google’s parent company, on the other hand was going elsewhere. Sales rose 1.3 percent. xAI is not publicly traded.

The announcement followed an agreement between the Pentagon and seven major technology companies – Google, Microsoft, Amazon Web Services, Nvidia, OpenAI, Reflection and SpaceX – using their AI systems in computer networks.

The Department of Defense said the deal would provide resources to help “increase military decisions in critical operational environments”.

Notably absent from the list is AI company Anthropic after public controversy and legal battles with the Trump administration over the ethics and safety of AI in the military.