Anthropic has discreetly erased numerous voluntary commitments from its website that it established with the Biden Administration in 2023 to foster safe and “trustworthy” AI.
These commitments included promises to exchange information on AI risk management between industry and government and to conduct research on AI bias and discrimination. Last week, these items were removed from Anthropic’s transparency hub, as reported by the watchdog group The Midas Project. However, other commitments aimed at addressing AI-generated image-based sexual abuse from the Biden era remain intact.
There has been no indication from Anthropic regarding this alteration, and the company did not respond promptly to inquiries for comments.
In July 2023, Anthropic, alongside firms like OpenAI, Google, Microsoft, Meta, and Inflection, announced its agreement to follow certain voluntary AI safety commitments proposed by the Biden Administration. These commitments encompassed conducting internal and external security assessments of AI systems prior to their release, investing in cybersecurity to safeguard sensitive AI data, and creating techniques for watermarking AI-generated content.
It is important to highlight that Anthropic had already implemented several of the practices mentioned in the commitments, which were not legally enforceable. The intention behind these commitments was to indicate the Biden Administration’s AI policy priorities ahead of a more comprehensive AI Executive Order that took effect a few months later.
Contrarily, the Trump Administration has signaled a markedly different direction concerning AI governance.
In January, President Trump rescinded the previously mentioned AI Executive Order, which directed the National Institute of Standards and Technology to develop guidelines to assist companies in identifying and rectifying flaws in their models, including biases. Trump’s supporters contended that the reporting requirements were excessively burdensome and effectively compelled companies to reveal their trade secrets.
Following the cancellation of the AI Executive Order, Trump issued a directive instructing federal agencies to encourage the development of AI “free from ideological bias,” aimed at fostering “human flourishing, economic competitiveness, and national security.” Notably, Trump’s order did not address combating AI discrimination, a crucial aspect of Biden’s initiative.
As the Midas Project reported through several posts on X, nothing in the commitments from the Biden Administration suggested that they were time-sensitive or dependent on the political affiliation of the sitting president. In November, following the elections, numerous AI companies confirmed that their commitments remained unchanged.
Anthropic is not the only corporation altering its public policies since Trump took office. OpenAI recently declared it would pursue “intellectual freedom… regardless of how challenging or controversial the subject may be,” while ensuring that its AI does not suppress specific viewpoints.
OpenAI also removed a page from its website that previously showcased the startup’s commitment to diversity, equity, and inclusion (DEI). These programs have faced criticism from the Trump Administration, prompting many companies to eliminate or significantly modify their DEI initiatives.
Several of Trump’s Silicon Valley advisors on AI, including Marc Andreessen, David Sacks, and Elon Musk, have alleged that firms like Google and OpenAI have engaged in censorship by restricting their AI chatbots’ responses. Labs such as OpenAI have denied that their policy adjustments stem from political pressure.
Both OpenAI and Anthropic are actively seeking government contracts or have already secured them.
Compiled by Techarena.au.
Fanpage: TechArena.au
Watch more about AI – Artificial Intelligence
