The US President, Donald Trump, orders against AI models that compromise truthfulness, accuracy, and historical facts. Trump signed the executive order in the ‘Winning the AI Race’ Summit organized in Washington, DC, on July 23, 2025. As per the order, the federal government is looking to transform the country into an AI export powerhouse by promoting the development of truthful AI models.
While AI models developed by other countries, such as China’s DeepSeek and Alibaba, India’s Perplexity, are receiving enhanced attention across the globe, the USA’s recent order seems nothing but a reaction to the increasing popularity of these models.
Moreover, by targeting the global competitors, the USA is doubtlessly aiming to preserve its AI dominance internationally.
Though Trump did not specify any AI model or developing company while announcing the ‘anti-woke AI’ principles but has mentioned incidents where AI models have been found manipulating historical facts in their outputs. The order targets and bans AI models that are not ‘ideologically neutral’ or ‘woke AI’ as the federal government had addressed it.
The changes came as a segment of the USA’s Winning the AI Race: America’s AI Action Plan.
The ‘Winning the AI Race’ initiative acknowledges the transformative power of innovations like AI and its potential to revolutionize several industries alongside offering new opportunities. Hence, the country aims to maintain its technological dominance unquestionably by leading the AI diplomacy and security globally.
Trump addresses the ideology of diversity, equity, and inclusion (DEI) as one of the most ‘pervasive and destructive’ dogmas that causes ‘distortion and suppression’ of facts. The misrepresentation of factual elements is mostly observed in ‘race or sex’ related outputs.
Alongside that, ‘manipulation with racial and sexual representation’, addition of critical race theory, unconscious bias, intersectionality, transgenderism, systematic racism, and race or sex related discrimination have become common in outputs of AI models with DEI considerations.
Significantly, just prior to the USA’s ‘anti-woke AI’ order, the Defense Department of the nation granted government contracts to firms, including Google, Anthropic, xAI, and OpenAI. The contract denoted awarding these companies with up to $200 million if they assist the agency in integrating advanced AI abilities and to ‘address critical national security challenges.’
Core Principles for LLM Development and Procurement:
Trump’s order to eliminate ‘woke AI’ highlights two key principles to consider while procuring and developing LLMs; these are-
- Truth-seeking: While prioritizing historical accuracy, objectivity, and scientific inquiry, alongside acknowledging the uncertainty of incomplete and contradictory information, LLMs must be truthful. The truthfulness has to be reflected throughout the outputs and responses made to the user prompts, seeking factual analysis.
- Ideological Neutrality: LLMs must be unbiased and neutral, offering non-manipulated outputs and avoiding adherence to ideologies such as DEI. Developers must ignore the intentional encoding of biased and ideological judgments into their LLMs’ responses. However, in case of specific prompting, judgements can have a space.
Nevertheless, these principles are subject to appropriate exceptions while using LLMs in national security systems.
Emerging LLM Concerns:
While referring to bias and manipulated outcomes of AI models, the ‘anti-woke AI’ order highlighted incidents that took place lately and raised concerns regarding the truthfulness of AI.
One example is when an AI model was found changing the race or sex of historical figures, such as the Founding Fathers and the Pope. Notably, the model was trained putting the ideology of DEI as a priority. On another occasion, an AI model denied generating images where white people celebrate their achievements while generating the same for people with other races and ethnicities.
Another case includes when an AI model affirmed that users must not ‘misgender’ others, even if it could prevent a nuclear disaster.
The new order will surely redefine the development and integration of AI in the USA. While aiming for safer and truthful development of AI, the USA is trying to maintain its ‘global technological dominance.’
Do not miss anything on AI and emerging tech trends. Visit HiTechNectar, and stay current!
Recommended For You: