By Julia Pastor
In February, Anthropic, a start-up specialising in generative AI, refused, on ethical grounds, to grant the Pentagon unlimited access to its AI model, Claude. The firm insisted on a number of red lines being respected regarding the use of its technology for mass surveillance of US citizens and autonomous weapons. As well as branding it “radical left” and “woke”, Donald Trump ordered the cancellation of all its contracts with the Department of Defence and included it on a list of companies posing a threat to national security.
Anthropic is a rare bird in the world of AI. Its founding manifesto is committed to AI that is safe, responsible and beneficial to humanity. Its CEO, Dario Amodei, left OpenAI in 2021 precisely because he believed it did not meet safety standards, and established his new company as a public benefit corporation: legally obliged to build models that respect human rights. Amodei has been open to dialogue with the White House, but has accused it of violating the First Amendment with its veto of Anthropic.
At the other end of the spectrum is Palantir, another of the sector’s major companies. Its manifesto, a 22-point text published in April on the social media platform X, has been unanimously described as an example of “techno-fascism”. In contrast to Amodei’s warning that the misuse of a technology as powerful as AI could derail democracy, Palantir’s CEO, Alex Karp, believes it is the primary weapon for ensuring the survival of the West. Bureaucracy and democracy hinder progress, and Silicon Valley has an almost moral duty to participate in the defence of the nation. The use of its technology by ICE to identify and track migrants in the US and its collaboration with the Israeli Ministry of Defence are well documented.
Amnesty International lists the US, China, Pakistan, Afghanistan, South Sudan and Tanzania among the countries that, by 2025, have used technology to implement authoritarian practices such as the unlawful surveillance of citizens to restrict the right to freedom of expression or suppress protests. Regulation of AI remains insufficient. For example, the AI Governance Guidelines developed in India were non-binding or very general.
