Lawsuit aside, White House holds talks with Anthropic over powerful AI tool Mythos

Nearly a month after declaring the company a 'supply chain risk,' the White House on Saturday held talks with the top executives of artificial intelligence firm Anthropic, describing the meeting as 'productive and constructive.'

The talks come at a time of growing concern over the company's powerful new AI system, Claude Mythos.

Anthropic CEO Dario Amodei, US Treasury Secretary Scott Bessent, and White House Chief of Staff Susie Wiles participated in the meeting.

According to a BBC report, discussions focused on possible cooperation as well as ways to manage the risks linked to advanced AI systems.

The talks are notable because they come just weeks after strong criticism from US President Donald Trump, who had previously ordered government agencies to stop using Anthropic's technology.

Concerns over powerful AI system

The meeting follows the recent release of Claude Mythos, an advanced AI tool that has raised both excitement and concern in the tech world. The system is currently available only to a limited number of companies but has already shown impressive abilities in cyber-security tasks.

According to Anthropic, Mythos can detect hidden bugs in old software and even identify ways to exploit them. Researchers have described the system as "strikingly capable," especially in areas that were once considered highly complex and human-driven.

These capabilities have caught the attention of US officials, who are increasingly aware of both the benefits and risks of such technology. The White House said that the meeting explored how to balance innovation with safety, including setting rules and safeguards for future use.

Legal dispute still unresolved

Despite the positive tone of the meeting, a legal battle between Anthropic and the US government is still ongoing. In March, the company sued the Department of Defence after being labelled a "supply chain risk," a designation that suggests its technology may not be secure enough for government use.

Anthropic argued that the label was unfair and politically motivated. It claimed that Defence Secretary Pete Hegseth imposed the restriction after the company refused to give the Pentagon unrestricted access to its AI systems. The firm said it was concerned about potential misuse, including mass surveillance and fully autonomous weapons.

A federal court in California largely supported Anthropic's position, but an appeals court later refused to temporarily lift the restriction.

Shift in tone from Washington

Even with the dispute, Anthropic's technology continues to be used by several government agencies. This suggests that officials see the tools as too important to ignore.

The recent meeting also signals a possible shift in tone from Washington. While the Trump administration had earlier criticised Anthropic sharply, the new discussions point to a more practical approach, where the US government finds it a compulsion to work with leading AI companies.