The Trump administration claims that Anthropic's AI model “endangers the lives of soldiers.” The skepticism of a judge
Pressed by the federal judges, the US administration defended on Tuesday its decision to designate the artificial intelligence startup Anthropic as a risk, which could “endanger the lives of soldiers”, reports AFP, taken by Agerpres.
The San Francisco-based AI company challenged in court its inclusion in early March by the US Department of Defense on a list of suppliers that pose “a risk to the supply chain”, a document that until now contained only foreign companies.
“I see no evidence”
The Trump administration took this step after Anthropic refused to lift certain restrictions on its Claude artificial intelligence model, preventing its use for mass surveillance of the US population and for lethal autonomous attacks.
Two legal proceedings are ongoing in this case, one before a court in San Francisco and the other before a federal appeals court in Washington.
The two courts issued conflicting rulings in summary proceedings: the San Francisco court temporarily suspended the ban on Anthropic, while the Washington court refused to do so.
During a hearing on Tuesday in the US capital, a representative of the Trump administration faced a series of questions, some marked by skepticism.
“I don't see any evidence provided by the Department of Justice to support the designation of (…) Anthropic” as a risk, said Karen Henderson, one of the three judges present, explaining that the facts and the applicable laws explain why the case is being heard by an appeals court and not a first instance.
According to the judge, the US administration, in this case, “dramatically exceeded its authority”.
Justice Department justifications
Trump administration spokeswoman Sharon Swingle argued that Anthropic's limitations on its AI model risked causing it to malfunction during a military operation, which “could have had catastrophic national security consequences and could have put the lives of soldiers at risk.”
“In order to raise awareness of this risk throughout the department and to be able to act quickly,” she added, “this designation was necessary.”
Anthropic attorney Kelly Dunbar expressed concern that “the Department of Defense is abusing this designation (as opposed to its original purpose) as a means of contractual pressure or retaliation against Anthropic for its disagreement” with the administration.
She reiterated that OpenAI's rival firm is not contesting the Trump administration's ability to terminate its contracts with it, but rather the process and arguments used.
A decision from the Court of Appeals in Washington is expected shortly, as the judges have already recognized that this sensitive case warrants an expedited procedure.
