In a subtle but significant shift, several major technology companies, including Anthropic, OpenAI, and Salesforce, have scaled back their commitments to artificial intelligence (AI) safety and diversity, equity, and inclusion (DEI) initiatives.
These moves come in the wake of President Donald Trump’s new executive orders, which promote a markedly different approach to AI governance and workplace diversity compared to the previous Biden administration.
Anthropic Withdraws AI Safety Pledges
Anthropic, a leading AI company, has quietly removed several voluntary commitments from its website that were made in conjunction with the Biden administration in 2023. These commitments aimed to promote safe and “trustworthy” AI, including pledges to share information on managing AI risks and conducting research on AI bias and discrimination.
Register for Tekedia Mini-MBA edition 17 (June 9 – Sept 6, 2025) today for early bird discounts. Do annual for access to Blucera.com.
Tekedia AI in Business Masterclass opens registrations.
Join Tekedia Capital Syndicate and co-invest in great global startups.
Register to become a better CEO or Director with Tekedia CEO & Director Program.
The deletions were flagged by AI watchdog group The Midas Project, which noted that while some Biden-era commitments, such as those addressing AI-generated image-based sexual abuse, remain, others tied to transparency and bias research were removed without notice.
Anthropic did not initially respond to requests for comment but later stated: “We remain committed to the voluntary AI commitments established under the Biden Administration. This progress and specific actions continue to be reflected in [our] transparency center within the content. To prevent further confusion, we will add a section directly citing where our progress aligns.”
The Biden administration’s AI safety commitments, announced in July 2023, involved industry heavyweights including OpenAI, Google, Microsoft, Meta, and Inflection. The companies had agreed to conduct internal and external security tests of AI systems, enhance cybersecurity measures, and develop methods for watermarking AI-generated content.
These voluntary commitments were not legally binding but were intended to set a standard for AI safety and ethical development. They also served as a precursor to Biden’s more formal AI Executive Order, which required federal agencies and AI developers to take significant steps to mitigate AI-related risks, particularly concerning bias and misinformation.
Trump’s Repeal of AI Safeguards
On Thursday, Trump repealed Biden’s AI Executive Order, marking a drastic shift in AI policy. The repealed order had instructed the National Institute of Standards and Technology (NIST) to create guidance for identifying and correcting biases in AI models. However, Trump’s new directive focuses on promoting AI development “free from ideological bias,” emphasizing “human flourishing, economic competitiveness, and national security.”
Trump’s order makes no mention of combating AI discrimination, a critical aspect of Biden’s initiative. Critics of Biden’s policies, many of whom are close allies of Trump, argued that the previous administration’s requirements were too burdensome and potentially forced companies to reveal trade secrets.
Marc Andreessen, David Sacks, and Elon Musk, some of Trump’s prominent Silicon Valley advisors, have accused companies like Google and OpenAI of AI censorship, alleging that their systems suppressed certain viewpoints. While OpenAI and others have denied any political influence on their policy changes, the recent rollbacks suggest a shift towards a more laissez-faire approach to AI development.
OpenAI and the End of ‘Ideological Bias’
OpenAI has also made changes to its public policies, removing references to diversity, equity, and inclusion (DEI) from its website. The company announced it would support “intellectual freedom … no matter how challenging or controversial a topic may be,” signaling a departure from previous commitments to mitigate AI bias and promote equitable outcomes.
This shift aligns with the Trump administration’s push against DEI programs, which it argues promote divisive “woke” ideologies. This is despite the belief that removing DEI frameworks from AI development could exacerbate the risk of discriminatory outcomes, particularly for marginalized communities.
Both OpenAI and Anthropic are actively pursuing government contracts, making their policy adjustments particularly relevant as they align more closely with the Trump administration’s priorities.
Salesforce Joins the Retreat from DEI Initiatives
In addition to changes in AI governance, companies are also backing away from DEI initiatives. Salesforce, a San Francisco-based customer relationship management giant, has removed language that once described diversity and inclusion as a “core value” in its latest financial disclosure.
The company’s move follows a series of executive orders from Trump aimed at eliminating DEI initiatives across both government and private sectors. Salesforce’s latest filing also omits previous language that tied executive compensation to employee diversity metrics.
“While we are not specifying representation goals, we remain committed to our core value of equality,” a Salesforce spokesperson told Bloomberg.
Salesforce CEO Marc Benioff, known for his vocal support of social issues, has yet to comment publicly on the rollback. The company’s decision mirrors similar actions by other tech giants, including Google, Meta, Amazon, Walmart, and Deloitte, which have all scaled back DEI programs since Trump took office.
A Turning Point for Tech Policy
The dismantling of Biden-era commitments to AI safety and DEI initiatives underlines a broader ideological shift in Silicon Valley. Under Biden, tech companies were encouraged to develop AI systems with robust safeguards against bias and to maintain inclusive hiring practices. Now, under Trump, the emphasis has shifted to promoting innovation and “ideological neutrality,” with less regulatory oversight.
The Midas Project pointed out that nothing in the Biden-era commitments suggested that the promises were time-bound or contingent on the sitting president’s party affiliation. This raises concerns about the stability of corporate commitments to ethical AI development and workplace equality.
AI ethics experts warn that these rollbacks could undermine trust in AI systems, particularly among communities that have historically been marginalized by biased technologies. The removal of DEI initiatives could also affect workplace diversity, potentially stalling progress on representation and inclusion.
However, many believe that Trump’s executive orders reflect a deregulatory stance aimed at reducing what his administration calls “onerous” requirements on tech companies.
The impact of these changes is expected to become clearer in the coming months, as companies adjust their policies and as the White House drafts its forthcoming AI action plan. This plan, expected within 180 days of Trump’s January executive order, will set the tone for the administration’s approach to AI innovation and public safety.