In the field of artificial intelligence, Anthropic's recent series of actions have attracted widespread attention. The well-known AI company was revealed to have removed several voluntary commitments jointly issued with the Biden administration from its official website, which were originally seen as Anthropic's positive statement on AI security and trustworthy artificial intelligence. This change was first discovered by the artificial intelligence supervisory agency "Midas Project" and quickly became the focus of the industry.
Anthropic's operation seems extremely low-key, and it can even be said to be carried out "quietly". The company did not take the initiative to announce this change, and chose to remain silent in the face of media requests for comment. This low-key approach makes the outside world full of speculation about the real intention behind it. Especially in the context of the current political climate change, this move seems even more intriguing.
Back in July 2023, Anthropic, together with technology giants such as OpenAI, Google, Microsoft, Meta and Inflection, publicly announced that it would respond to the Biden administration's call and voluntarily abide by a series of artificial intelligence security commitments. These commitments cover key measures such as rigorous internal and external security testing before the release of AI systems, vigorously investing in cybersecurity to protect sensitive AI data, and research and development of watermarks for AI-generated content.
Although these commitments are not legally binding, they are given a deeper political significance in the Biden administration's strategic layout. These promises are seen as an important signal for the White House to send out priority directions for its AI policies ahead of the upcoming more comprehensive AI executive order. However, with the Trump administration taking office, this policy trend has changed significantly.
In January this year, the Trump administration decisively abolished the Biden administration's executive order for artificial intelligence. The order was originally intended to guide the National Institute of Standards and Technology to develop industry guidelines to assist businesses in identifying and correcting flaws in their models, including highly-watched bias issues. The Trump administration believes that the cumbersome reporting requirements in the Biden administration's orders actually force companies to leak trade secrets, which is something they cannot accept.
Immediately afterwards, Trump quickly signed a new executive order directing federal agencies to actively promote the development of AI technology "unaffected by ideological bias", with the goal directly aiming at promoting "human prosperity, economic competitiveness and national security." It is worth noting that Trump's new order does not mention the core principle of "fighting AI discrimination", which is precisely the key content of the Biden administration's initiative.
As the "Midas Project" pointed out on social media platforms, the AI security commitment in the Biden administration era never implies that it has time limits, nor does it have any connection with the partisanship of the current president. However, in just a few months since the Trump administration came to power, Anthropic is not the only technology company to make adjustments at the public policy level. OpenAI recently announced that it will embrace "leisure of knowledge" and emphasized that no matter how sensitive or controversial the topic is, it should be open-minded.
OpenAI even quietly removed a page on its official website dedicated to articulating the company's commitments in diversity, equity and inclusion (DEI). These DEI plans have been severely criticized by the Trump administration, and have directly led to many companies stopping or significantly adjusting their original DEI plans. Many Silicon Valley artificial intelligence consultants, including Mark Anderson, David Sachs and Elon Musk, have publicly accused tech giants including Google and OpenAI of implementing so-called "artificial intelligence censorship" by artificially restricting the answers of their AI chatbots.
A background that cannot be ignored is that OpenAI and Anthropic, two leading companies in the AI field, have long been in the track of competing for government contracts and have performed quite positively. Just as public opinion continues to ferment, Anthropic finally broke the silence and issued a statement to the outside world trying to calm the controversy: "We remain firmly committed to the voluntary AI commitments established by the Biden administration. The progress and specific actions of these commitments are reflected in the content of [our] Transparency Center. To avoid further confusion, we will add a special section to directly quote our progress."
However, can this belated statement completely answer external doubts? Is Anthropic's "promise withdrawal" incident just a misunderstanding, or a strategic turn after careful consideration? In the changing political and business game, every move of AI giants deserves our high attention.