Without really knowing why, suddenly, the American start-up Anthropic, known for its chatbot Claude, published a press release reaffirming that it was aligned with the Trump administration’s objectives in terms of artificial intelligence.
“Our long-standing position is that managing the social impacts of AI should be a matter of measurement rather than policy. I am convinced that Anthropic, the administration and leaders across the political spectrum want the same thing: to ensure that the power of AI benefits the American people and that the United States progresses and consolidates its leadership in the development of AI,” said the head of the start-up, Dario Amodei.
Crisis management and “woke” position
However, this post doesn’t come out of nowhere. If Dario Amodei considered it necessary to make this “clarification”, it is to respond to a “resurgence of inaccurate statements” about the political positions of his company and, therefore, to appease Donald Trump and his administration.
It all began with the publication by Jack Clark, Anthropic co-founder and public policy director, of a speech titled “Children in the Dark,” which he gave in Berkeley at an AI conference called The Curve.
In this speech, he addresses in particular the potential dangers posed by artificial intelligence. A position that was not appreciated by one David Sacks, chairman of the United States Presidential Committee of Advisors on Science and Technology.
The venture capitalist also spoke about it in
Everything could have stopped there and the world would not have been angry, but the co-founder of Linkedin (a big donor to the Democratic Party, as pointed out by the outlet 404, which chronicles the skirmish with talent) joined the dance in a long thread, where he explains that it is important, especially in the AI industry, to support “good people.” This is anthropic.
Which obviously provoked a reaction from the Trumpist AI advisor, in these terms: “The real problem (lies) in Anthropic’s plan to surreptitiously introduce “woke” AI regulation and other regulations in Democratic states like California.” I woke up. The preliminary deadline for ostracism was published, requiring an immediate countermeasure to appease possible presidential anger.
Towards political neutrality
That is why Darío Amodei was forced to publish this declaration of principles that smells like a knee on the ground. For this reason, he notably stated that Anthropic’s AI models are not “particularly biased” when it comes to politics.
To back up his comments, he cited two studies, including one published in January by the Manhattan Institute that found those from other companies to be more biased than its then-main model, the Claude Sonnet 3.5.
To take a further step back, these statements also come a few months after Donald Trump presented his action plan on AI, which aims, among other things, to develop AI without “ideological biases.”
A shared vision
Anthropic, through its manager, recalls that it welcomed this action plan when it was announced last July. “We support the president’s efforts to expand America’s energy supply to win the AI race,” said Dario Amodei. He also took the opportunity to mention the summit on AI and energy in which he participated with Donald Trump that same month.
Among other initiatives of his startup, the CEO also mentions that he recently offered his AI models to the US government for a symbolic dollar or that he hired political experts from the Republican and Democratic parties. Anthropic also formed an advisory board made up of former Trump administration officials.
“We are willing to work in good faith with anyone, regardless of their political affiliation,” said Dario Amodei. This, with the aim of making the vision that the startup shares with the American president a reality. That is, “try to maximize the positive aspects and minimize the negative aspects of AI, since these systems are neither all good nor all bad.
Source: BFM TV
