In case you missed it, the U.S. Government has recently been clashing with one of the biggest AI companies over how it is using their technology.
At the heart of the argument is how AI is used:
- To surveil populations.
- In military operations.
Here’s how it has unfolded.
AI race
Last year, U.S. President Donald Trump detailed the country’s ‘AI Action Plan’ and declared that the country is “in a race to achieve global dominance in artificial intelligence”.
The President compared this to the space race of the 1960s.
Part of his plan was the need for AI to be “aggressively” adopted within its military operations, which he said could transform the country’s “warfighting”.
Anthropic
Anthropic was the first generative AI company to have its models deployed on the U.S. government’s classified networks.
It is one of the biggest AI companies in the world, best known for developing Claude (an AI chatbot and a series of other large language models).
Last year, it was awarded a $US200 million two-year contract with the U.S. Department of War. The partnership was designed to “advance U.S. national security.”
Last week, Anthropic made public their request for their contract to now include two exceptions to the Department of War’s use of its models.
Those exceptions were that their technologies should not be used for:
- Mass domestic surveillance. The company said today’s AI could assemble “a comprehensive picture of any person’s life – automatically and at massive scale”. Anthropic said it doesn’t want its models used like that, because it believes it would “violate fundamental rights”.
- Fully autonomous weapons. These are weapons that don’t require humans to identify and shoot at targets. Anthropic said that whilst this could be “critical” in the future, the technology is not reliable enough yet. It said this needs to be “deployed with proper guardrails, which don’t exist today”.
Anthropic said these two exceptions, if allowed, would “undermine, rather than defend, democratic values”.
Your contribution ensures The Daily Aus can continue doing the work you love.
I want to give $10 monthly to help TDA keep doing the news.
Your details:
U.S. response
The U.S. Government disagreed. It argued it should be able to use Anthropic’s models however it wants, as long as it’s legal. (Anthropic argues the law hasn’t caught up yet).
The government negotiations were being led by the U.S. Defense Secretary Pete Hegseth, who is the head of the Department of War.
Last week, he imposed a deadline for Anthropic to comply to its demands. That deadline was last Friday at 5:01pm.
Anthropic did not comply with that deadline, and refused to change its demands.
At 5:14pm last Friday, Hegseth issued a statement announcing he had designated Anthropic as a national security risk.
In that statement, Hegseth said the President and the American people “will determine the destiny of our armed forces, not unelected tech executives.”
He added: “America’s warfighters will never be held hostage by the ideological whims of Big Tech.”
Trump responded in a post to Truth Social, saying he had directed the Government to stop using "radical left, woke" Anthropic.
OpenAI
Hours after the deadline passed, OpenAI announced it had reached an agreement with the Department of War, which would replace Anthropic’s deal.
OpenAI is the maker of ChatGPT, the most widely used AI chatbot.
The agreement said the technology could be used for any lawful purpose.
In a post to X, CEO of OpenAI Sam Altman said the government had “displayed a deep respect for safety” in their interactions.
However, OpenAI then amended the contract. Sam Altman said a few days later that he “shouldn’t have rushed” the deal and said his initial statement “just looked opportunistic and sloppy”.
The new amendments include no use of its technology for mass domestic surveillance, and no use of its technology to direct autonomous weapons systems. (These were Anthropic’s initial requests).
OpenAI says it has requested that Anthropic now be offered the same terms.
