White House officials are drafting an executive order that would penalize tech businesses that utilize what they perceive to be “woke” artificial intelligence models. This is the administration’s latest attempt to attack inclusion, equity, and diversity programs.
In an effort to counteract what administration officials perceive as unduly liberal AI models, the directive would require AI businesses that get federal contracts to be impartial and politically neutral, according to the sources.
Recent years have seen the proliferation of AI chatbots, such as Google’s Gemini, which some conservatives claim are politically liberal. Businesses have faced criticism for incorrectly portraying historical characters as persons of race in their imagery or for certain rapid replies that have angered customers. When Google’s Gemini AI assistant displayed a black George Washington and ethnically diverse Nazis last year, the company came under heat.
The trend has caused concern for two of Trump’s top tech advisors, Sriram Krishnan, senior White House policy adviser for AI, and AI Czar David Sacks, according to the sources.
Since almost all of the biggest tech companies are fighting for the federal government to adopt their AI capabilities, the directive might have a significant influence and make developers very cautious about how they build their models.
In order to train their models, AI businesses scrape a large amount of data from published sources and the internet. Occasionally, the models “hallucinate,” producing unexpected or incorrect answers. This might make it hard to identify the source of apparent bias or inaccuracy.
According to the sources, President Trump’s strategy for defeating China in the AI race would be outlined in a number of executive orders that are anticipated to be issued next week. At that moment, Trump is anticipated to discuss his AI “action plan.”
According to the sources, another order is anticipated to use institutions such as the U.S. Export-Import Bank to encourage exports of American chips and artificial intelligence tools. The government has made it a point to send technology to U.S.-friendly nations so that they may use American artificial intelligence (AI) rather than Chinese goods. Sacks and Krishnan have led an initiative to increase Nvidia’s exports of high-performance AI processors to nations like the United Arab Emirates.
It’s also anticipated that attention will be paid to expediting the permitting process for data centers required for model training and increasing the energy generation for such data centers. Tens of billions of dollars have been invested in such kinds of initiatives, as Trump, Sacks, and other government officials celebrated earlier this week in Pittsburgh.
A request for comment from the White House was not immediately answered. A representative for the Office of Science and Technology Policy, which is in charge of part of the AI research, chose not to respond.
The directive aimed at “woke” AI is considered to be one of the most contentious measures, as some in the tech industry may perceive it as favoring developers.
Elon Musk’s xAI, for example, has prioritized political neutrality and “anti-woke”. However, it has came under fire for displaying antisemitic propaganda celebrating Hitler via its Grok chatbot. Musk and Sacks are good friends and ex-colleagues. Musk just quit the government following a public spat with Trump.
Some firms have attempted to construct AI models without constraints in order to circumvent what they perceive as censorship.
After recruiting a number of high-ranking Biden administration personnel, Anthropic, a highly regarded AI developer founded by former OpenAI executives, may likewise be perceived as more liberal. According to several of the sources, Sacks and other officials have criticized Anthropic for opposing a number of the administration’s positions on chip exports and AI rules. According to Anthropic, there is no political motivation behind its actions.
Since the beginning of his first presidency, Trump has attacked conservatives online for what he considers to be bias. Using executive orders and other measures, he has waged a war against DEI programs throughout his second term, aiming to destroy the federal government’s and businesses’ associated activities.
