On July 23, 2025, President Donald J. Trump signed an executive order titled "Preventing Woke AI in the Federal Government," which prohibits federal agencies from procuring artificial intelligence models deemed to contain ideological biases. The directive aims to ensure that AI systems used by the government adhere to principles of truthfulness and ideological neutrality. This move reflects a significant shift in federal AI procurement policy, as highlighted by investor David Sacks, who stated on social media, "the federal government will not be buying WokeAI."
The executive order mandates that federal agencies procure only Large Language Models (LLMs) developed in accordance with newly defined "Unbiased AI Principles." These principles require AI systems to be "truth-seeking" and "ideologically neutral," explicitly stating that developers should not intentionally encode partisan or ideological judgments into an LLM’s outputs unless prompted by the end user. The Office of Management and Budget (OMB), in consultation with other federal leaders, is tasked with issuing guidance within 120 days to implement these procurement standards.
The Trump administration frames this order as essential for maintaining the integrity of information and services provided by the government, asserting that AI models should not sacrifice accuracy for ideological agendas. President Trump stated that "The American people do not want woke Marxist lunacy in the AI models." This initiative is part of a broader campaign to dismantle Diversity, Equity, and Inclusion (DEI) policies across federal departments, with the administration emphasizing a return to meritocracy and objective standards. David Sacks, the White House AI and crypto czar, was instrumental in shaping this policy.
The order carries significant implications for AI companies seeking federal contracts, as future business could hinge on their models complying with the new unbiased principles. Contracts will include terms requiring adherence, with provisions for termination and vendor accountability for decommissioning costs in cases of non-compliance. While the order emphasizes hesitancy to regulate the private marketplace, it asserts the government's obligation not to procure models that compromise truthfulness. Critics, however, argue that the definition of "woke AI" remains vague and could lead to subjective enforcement, potentially stifling innovation.
This executive action also aligns with the administration's broader "AI Action Plan," which seeks to cement the United States' global dominance in artificial intelligence. By ensuring AI systems are free from perceived ideological bias, the government aims to foster trustworthy AI while accelerating innovation. The policy underscores a strategic focus on national security and economic competitiveness in the evolving global AI landscape, particularly in the context of competition with other AI powers.