Skip to Content

Welcome to the age of AI-driven discrimination

'AI accountability is out and tech infrastructure is in,' a policy expert told Gazetteer SF

Silicon Valley is all in on artificial intelligence technologies, but there’s currently “a huge amount of snake oil,” “crummy AI” systems, and “shoddy” products, Kit Walsh, the Electronic Frontier Foundation’s director of AI, told Gazetteer SF.

That’s why President Trump’s decision this week to rescind the Biden-era executive order pushing for safeguards in AI technology “is a big step in the wrong direction,” Walsh said. “And when coupled with the sort of enthusiasm for proven-to-be-faulty AI technologies that a lot of the Trump advisors have shown, it’s concerning.”

Biden’s 2023 order required agencies in the executive branch, like the Department of Defense, Department of Education, and Department of Housing and Urban Development, to implement safeguards against unintended consequences related to AI, noting such precautions were especially important in areas like healthcare, education, and housing. 

Walsh said she fears we’ll see more “use of a junk AI” that either wastes taxpayer money, or in the worst case scenario, automates harms, and deprives the public of due process and access to government services.

“When you have a crummy AI system making those decisions, it deflects accountability and often creates new forms of discrimination,” Walsh said. “Or, the AI is trained on a historical data set that already reflects unfair and discriminatory decisions, and then you sort of institutionalize it while at the same time making it appear neutral because it's coming out of a computer.”

While Biden’s executive order did have some reporting requirements for AI companies, there were no penalties to enforce them, according to Walsh. Its primary role was to push for smarter use of AI within the federal government. Still, it served as “a value statement when it came to the private use of AI,” he said.

Trump, on the other hand, has argued the executive order stifles innovation. 

Shortly after rescinding Biden’s order, Trump teamed up with tech elite like OpenAI CEO Sam Altman and Oracle Executive Chairman Larry Ellison to announce the federally-supported Stargate Project, a new company that plans to invest $500 billion to “secure American leadership in AI,” according to a blog post about the project. 

That likely includes building more water- and energy-hungry data centers on U.S. soil, which comes with a massive environmental footprint (including, Trump has said, the expanded use of coal for power generation).  

As part of the press conference announcing the Stargate Project, Trump spoke about the importance of the joint venture between OpenAI, Oracle, and SoftBank amid competition from China, saying he wants to keep AI innovation in the U.S.

“Less than a week into the administration, it’s clear AI accountability is out and tech infrastructure is in,” Brian J. Chen, a policy director at nonprofit research organization Data & Society, told Gazetteer. “Trump’s priorities are all about national security and winning the supposed AI Cold War against China.” 

While Biden’s executive order was imperfect, Chen said, with certain carve-outs and waivers “that were always going to undermine its effectiveness,” it at least “represented a policy approach that understands that AI presents harms to people in society and that the government should lay down guardrails to protect people’s rights.” 

Now, that’s all out the window. It’s fair to say we no longer live in a society where AI safety is a priority for our nation’s leaders. 

And as the focus shifts from AI safety and accountability to building the infrastructure needed for AI dominance, the winners will be private investors and tech companies, Chen said. The rest of us are likely to be the losers.

“We’re likely to see these massive industrial buildouts that have little relationship to people’s everyday needs,” he said. 

Policy experts told Gazetteer that states will need to take the lead on regulating the private use of AI, if the feds are unwilling to mitigate the technology’s harms. EFF, for example, is in constant communication with state legislators, and Walsh said there are three AI-related proposals California lawmakers are likely to propose in the upcoming legislative session. 

“If political will doesn’t exist at the federal level, we need to be focused on the area where there is political will to implement those policies,” Chen said.

Stay in touch

Sign up for our free newsletter

More from Gazetteer SF

Ahead of Nintendo’s downtown arrival, a look back at the beloved PlayStation store

Perhaps the 90's-era Sony Metreon was ahead of its time, but now Nintendo has a chance to recapture some of the magic it left behind

April 18, 2025

Mr. Tan goes to Washington… and the battle over ‘Little Tech’ is heating up

Silicon Valley leaders, regulators, and elected officials came together earlier this month to debate the future of small startups, but the issue is more divisive than it seems

April 17, 2025

On the heels of the WNBA Draft, Valkyries super-fans are prepared to be the loudest in the league

The Valkyries made three key picks in the WNBA draft ahead of the inaugural season

April 15, 2025

One Montgomery is a columned monument to glory days of banking in SF

Downtown pile features Sierra granite, marble staircase, and unique copper canopy

April 15, 2025