The Trump administration wants to remove ‘ideological bias’ and fairness in the development of AI technology at a critical juncture.
The National Institute of Standards and Technology has instructed scientists that work with the US Artificial Intelligence Safety Institute to avoid considering “AI safety”, “responsible AI”, and “AI fairness” in the skills it expects of members, according to a report in Wired.
The AI Safety Institute’s new cooperative research and development agreement, released in March, wants to “reduce ideological biases” and counters previous agreements that encouraged researchers to identify and fix AI model biases related to gender, race, age or wealth inequality.
The new directive also removes mention of developing tools that would track and identify misinformation and deep fakes, while urging one working group to develop testing tools “to expand America’s global AI position”.
A researcher at the AI Safety Institute, who spoke to Wired under the condition of anonymity, believes the new instructions could harm regular users by allowing algorithms to discriminate on the basis of race and socioeconomic background.
The researcher said: “Unless you’re a tech billionaire, this is going to lead to a worse future for you and the people you care about. Expect AI to be unfair, discriminatory, unsafe, and deployed irresponsibly.”
Why this matters
Ethical AI has become a hot topic in recent years with the emergence of OpenAI’s Chat GPT, Meta’s Llama and Google’s Gemini.
Last year, a UNESCO study revealed worrying tendencies in Large Language models (LLM) to produce gender bias, as well as homophobia and racial stereotyping.
The study, which assessed OpenAI’s GPT-3.5 and GPT-2, and Meta’s Llama 2, found women were described as working in domestic roles far more often than men – four times as often by one model – and were frequently associated with words like ‘home’, ‘family’ and ‘children’; while male names were linked to ‘business’, ‘executive’, ‘salary’, and ‘career’.
Donald Trump’s closest advisor and largest commercial backer, Elon Musk, has previously labelled OpenAI as “woke” and Gemini as “racist”. Musk runs a competing AI company, xAI, and has already used his position as the head of the so called Department of Government Efficiency (DOGE) to slash government agencies that are investigating his business interests, while cancelling contracts to companies that rival his own.
Ashwini K.P, the UN Special Rapporteur on racial discrimination, xenophobia and intolerance, released a report that found that the use of artificial intelligence perpetuated racial discrimination, particularly in fields such as law enforcement, healthcare and education.
“In short, bias from the past leads to bias in the future,” she said, calling for stronger regulation to provide balance and safeguards.
Instead, it appears that the Trump administration is pushing Silicon Valley and the future of AI technology into a different direction.
