(New York) start-up of artificial intelligence (AI) Anthropic on Monday sued several entities of the American administration before a federal court in California to obtain the lifting of sanctions which it considers excessive and which could threaten its activity.
Published at
The Department of Defense last week added Anthropic to its list of companies presenting a “national security risk to supplies” in retaliation for the company’s refusal to lift restrictions on the use of its AI.
The filing of this summons, confirmed Monday by Anthropic to AFP, is only a formal step for the group which had already announced on Thursday its intention to take the matter to court. Asked by AFP, the department did not respond immediately.
The addition to the list of “at risk” companies came after the declaration of Donald Trump, who ordered, at the end of February, all government departments to “immediately cease all use of Anthropic’s technology”.
The American president thus intended to sanction OpenAI’s competitor for its refusal to see its AI models used for mass surveillance of populations or the automation of deadly attacks by the American army.
“These measures are unprecedented and illegal,” argues Anthropic in the summons document, consulted by AFP. “The (US) Constitution does not authorize the government to exercise its enormous power to punish a corporation for using its freedom of expression. »
These decisions are “arbitrary”, “a caprice” and “an abuse of discretionary power” of the American government, underlines the start-up from San Francisco. According to her, the government “seeks to destroy” Anthropic.
PHOTO BHAWIKA CHHABRA, REUTERS ARCHIVES
Anthropic boss Dario Amodei
The young company reveals that after the Trump administration’s announcements, it was contacted by “numerous partners and customers, remote (cloud) computing providers and investors”, who expressed their “confusion”.
Some customers also asked Anthropic under what conditions they could terminate their contract.
The group is not only contesting its inclusion on the list of “at risk” suppliers but also the conditions for the termination of contracts binding it to the government, considered abusive.
The list so far only included foreign companies, including the Chinese equipment manufacturer Huawei and the Russian antivirus software specialist Kaspersky.
It generally applies to entities whose products are deemed to threaten the national security of the United States.
The “at risk” classification theoretically prevents any other US federal government supplier from contracting with the company in question.
But Anthropic boss Dario Amodei revealed Thursday that the terms of the letter sent by the Department of Defense were “narrow in scope” and only prohibited the use of its products within the framework of their contracts with the Department of Defense.
Furthermore, Google, Microsoft and Amazon explained that Anthropic’s AI models available on their platforms would remain available for all uses other than those related to the Ministry of Defense.
