White House Aims to Toughen Rules for AI Partnerships

These guidelines would require AI companies to permit “any lawful” use of their models, the Financial Times (FT) reported Friday. 

The report, citing a draft of the guidelines seen by the news outlet, says the rules stipulate that companies wishing to work with the government give the U.S. an irreversible license to use their systems for all legal purposes.

The guidance from the U.S. General Services Administration would cover civilian contracts and is part of a government-wide effort to bolster procurement of AI services, the report added.

A source familiar with the matter said the rules are similar in principle to measures the Pentagon is weighing for military contracts.

The planned rules come as the government is engaged in a legal battle with AI startup Anthropic over its proposed use of that company’s technology. 

The Defense Department said late last month that it was exiting its contract with Anthropic after that company expressed concerns about the military using its systems for domestic surveillance and autonomous weapons.

The White House also took the unusual measure of naming Anthropic a supply-chain risk, which means companies that want to do business with the U.S. government are barred from working with Anthropic.

FT noted Defense Secretary Pete Hegseth said the company’s “true objective” was “to seize veto power over the operational decisions of the United States military.” Anthropic has said it will go to court to challenge the supply-chain risk designation.

Meanwhile, Anthropic rival OpenAI has inked its own agreement with the Pentagon. That company was reportedly able to get the military to agree to the “red lines” behind the Anthropic rift: mass domestic surveillance or autonomous weapons.

In other AI news, new PYMNTS Intelligence research shows that many product leaders now view gen AI not just as a tool for creativity and automation, but as a way to tighten oversight, boost compliance and lessen operational friction. 

“That is a big change from a year earlier, when expectations were meaningfully lower and many organizations were still deciding whether the technology belonged beyond pilot projects,” PYMNTS wrote last week.

The research showed that 98% of product leaders expect gen AI to improve internal workflows within three years, compared to 70% two years ago. Ninety-five percent said they expect more accurate decision-making, up from 67%; 83% are eyeing stronger data security, up from 50%.

Provider preferences are divided, with half of tech firms naming OpenAI as the leading provider, around 30% of goods firms choosing Google, and 24% of services firms favoring Microsoft.

For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.