Silicon Valley says Trump’s ‘woke AI’ crackdown – Business News
Big Tech is pushing back towards the Trump administration’s effort to stamp out “woke” artificial intelligence — in spite of a latest bombshell report discovering strongly left-leaning bias amongst main AI chatbots.
A software program industry commerce group representing Microsoft, OpenAI, Oracle, Salesforce and IBM says the administration is setting AI corporations up to fulfill necessities no one can reliably fulfill — a transfer they warn may persuade some AI distributors that doing business with Washington isn’t definitely worth the headache.
The group, referred to as the Business Software Alliance, is urging the federal General Services Administration to rewrite proposed contract guidelines that might require AI systems offered to the federal government to adjust to the administration’s “Unbiased AI Principles.”
President Trump has ordered federal businesses to favor AI systems which can be “truth-seeking” and “ideologically neutral.” AFP through Getty Images
The commerce group mentioned the draft guidelines, launched for public remark final week, are “not feasible to implement,” depend on imprecise requirements and will expose contractors to legal and financial dangers.
The transfer got here weeks after an in-depth evaluation by the Washington Post discovered lefty bias amongst main AI chatbots on points from DEI and homosexual conversion to marketing campaign finance and defunding the police.
News of the BSA submitting was first reported by NOTUS.
“Our comments below identify the unintended consequences of the draft clause, which will thwart this important objective and limit federal access to cutting-edge AI products and services,” the BSA wrote in feedback submitted July 23.
The struggle stems from President Trump’s govt order directing federal businesses to buy AI fashions which can be “truth-seeking” and “ideologically neutral.” The Office of Management and Budget later turned these rules into procurement steerage that GSA is now attempting to put in writing into authorities contracts.
If the rule takes impact, corporations promoting AI to Uncle Sam must certify their systems meet these requirements, undergo authorities testing and risk dropping contracts if officers conclude they don’t.
Anthropic grew to become a flashpoint in Trump’s struggle over AI guardrails after clashing with the Pentagon over army makes use of of its technology. NurPhoto through Getty Images
BSA argues that’s simpler mentioned than performed.
The group says that’s as a result of immediately’s AI doesn’t work like a calculator. Large language fashions generate probabilistic responses reasonably than assured factual solutions, making good accuracy unimaginable.
“Current technical realities do not enable guarantees of factual accuracy, as they are inherently probabilistic,” the group wrote. It added that “‘nonpartisan’ and ‘neutral’ are not sufficiently clear guideposts that can be translated into infallible technical measures.”
Instead of requiring compliance, the group urged GSA to undertake a “commercially reasonable efforts” commonplace, arguing that builders shouldn’t be anticipated to guarantee outcomes that stay exterior their technical control.
The submitting additionally says many software program corporations couldn’t adjust to some of the proposed guidelines as a result of they depend on AI fashions constructed by third events, making it impractical to “flow down” federal contract necessities to corporations they don’t control.
The group additionally warned that guidelines protecting open-source software program, mental property, knowledge portability and authorities testing may make AI corporations suppose twice about promoting their merchandise to Washington.
BSA mentioned the principles may finish up driving AI corporations away from federal contracts — undercutting the administration’s own push to hurry AI adoption throughout authorities.
Tech corporations say Washington is asking AI systems to fulfill requirements they can not reliably guarantee. NurPhoto through Getty Images
Ryan Frazee, a authorities contracts accomplice at Mayer Brown, instructed The Post that the proposal asks contractors to fulfill requirements which can be troublesome to outline, a lot much less implement.
“It’s a very vague and subjective standard. Whether an AI model’s response to a specific question is biased is going to be in the eye of the beholder,” he mentioned.
Frazee added that the proposal additionally misunderstands how immediately’s AI systems work.
“They’re not fact retrievers; they’re probabilistic models,” he mentioned.
“The requirement that there needs to be a factually accurate response with that kind of system is unrealistic.”
The Washington Post report just lately took main AI chatbots to activity, discovering OpenAI’s GPT-5.5 answered almost each query “exclusively with left-leaning arguments” when requested about 29 hot-button points. Anthropic’s Claude Opus 4.8 supplied simply a lefty argument 43% of the time, gave “both-sides” solutions 47% of the time — and by no means served up simply a right-leaning reply, in keeping with the evaluation.
The proposed new guidelines got here within the wake of Trump’s conflict with Anthropic, which he branded a “radical left, woke” company after it resisted giving the Pentagon unrestricted use of its AI systems.
The president ordered federal businesses to stop utilizing Anthropic’s technology, whereas the Pentagon designated the company a supply-chain risk.
Anthropic CEO Dario Amodei has discovered his company in President Trump’s crosshairs after the president branded the AI firm “radical left” and “woke” over its limits on army use of its technology. REUTERS
The dispute centered on Anthropic’s effort to keep its fashions from getting used for totally autonomous weapons or mass home surveillance.
Anthropic later sued, calling the federal government’s actions retaliatory, and a federal choose questioned whether or not the administration had justified the blacklisting.
Trump later mentioned he no longer considered the company as a national security risk.
“GSA is in the process of establishing acquisition requirements to protect government data in LLM systems, and we welcome further stakeholder input on the rule,” the GSA mentioned in a assertion supplied to The Post.
