Tech Giants Face New AI Showdown: DEI Efforts Under Fire as ‘Woke AI’ Becomes Political Battleground

 

Tech Giants Face New Scrutiny Over 'Woke AI' and DEI Efforts in Artificial Intelligence

CAMBRIDGE, Mass. – Major tech firms, already stepping back from internal diversity, equity, and inclusion (DEI) initiatives, might soon deal with another wave of challenges. This time, it's about how they've woven DEI principles into their AI technologies.

Under the current White House and a Congress controlled by Republicans, the spotlight has shifted from fixing harmful biases in algorithms to tackling what they call "woke AI." Previous attempts to promote fairness in AI creation and prevent biased results are now under the microscope. Last month, the House Judiciary Committee issued subpoenas to big players like Amazon, Google, Meta, Microsoft, OpenAI, and 10 others, probing these very efforts.

Meanwhile, the U.S. Commerce Department's standards division has scrubbed references to AI fairness, safety, and "responsible AI" from its calls for researcher partnerships. Instead, it's pushing experts to tackle "ideological bias" to boost human progress and economic strength, based on a document reviewed by The Associated Press.

Tech employees are no strangers to these sudden policy swings from Washington that shake up their daily work.

Experts Sound Alarm on Shifting AI Priorities

This latest pivot is worrying specialists in the field, such as Harvard sociologist Ellis Monk. A few years back, Google reached out to him to enhance inclusivity in their AI tools.

At the time, the industry was grappling with issues in computer vision – the AI tech that teaches machines to interpret images. It promised big business gains but carried over old prejudices from camera tech that often misrepresented people with darker skin tones.

"People with Black or darker skin would show up in photos looking off or silly," explained Monk, who studies colorism – discrimination tied to skin shade and features.

Google integrated Monk's innovative skin tone scale, which better captured human diversity in AI imaging. It swapped out an outdated system meant for white patients in dermatology.

"People loved the updates – feedback was overwhelmingly positive," Monk noted.

Today, he's questioning if similar projects will persist. He feels his scale is safe since it's embedded in countless products, from smartphones and games to AI art creators. Still, he and peers fear the current climate could freeze out new ideas and funding aimed at making tech fairer for all.

"Google aims for products that serve everyone, whether in India, China, or Africa. That's somewhat protected from DEI backlash," Monk said. "But yeah, funding for these could drop when politics heat up and everyone's rushing to launch."

Trump's Influence on AI Development and Bias Concerns

President Trump has slashed hundreds of grants in science, tech, and health that touch on DEI topics. The impact on commercial AI like chatbots is subtler. In the probes, Rep. Jim Jordan (R-Ohio), who leads the Judiciary Committee, is digging into whether the Biden team pressured companies to suppress free speech.

At a recent Texas gathering, Michael Kratsios, head of the White House Office of Science and Technology Policy, criticized Biden's AI approach for "fostering divisions and pushing equity-driven redistribution."

The administration wouldn't let Kratsios speak directly but pointed to examples, like a Biden policy snippet warning that unchecked AI could "worsen unfair outcomes for people and groups."

Long before Biden, studies and stories highlighted AI bias dangers. One report found self-driving cars struggle to spot darker-skinned walkers, raising accident risks. Another showed AI image tools depicting surgeons as white men 98% of the time – way off real-world stats, even in a male-heavy profession.

Phone unlock tech often bungled Asian faces. U.S. police made wrongful arrests of Black men due to faulty facial recognition. And remember a decade ago when Google's photo app tagged two Black individuals as "gorillas"?

Even during Trump's first term, federal experts in 2019 noted facial recognition inconsistencies across race, gender, and age.

Biden's win sped up some companies' push for AI equity. Then ChatGPT from OpenAI hit in 2022, kicking off a rush in AI for writing and visuals, forcing rivals like Google to speed up.

The Google Gemini Fiasco and Rise of 'Woke AI' Criticism

Enter Google's Gemini chatbot – its bungled launch last year became the poster child for "woke AI" that critics on the right want to dismantle. Untamed AI image generators often spit out stereotypes from their training data.

Google's version leaned toward lighter skin, men, and young women in job depictions, per their own studies. To fix it, they added safeguards before launch over a year ago. But it backfired, inserting diverse figures into wrong historical scenes – like showing America's Founding Fathers as Black, Asian, or Native American men.

Google said sorry and paused the tool, but it fueled conservative outrage.

At a February AI event in Paris, with Google CEO Sundar Pichai close by, Vice President JD Vance slammed "ahistorical agendas in AI," citing Gemini's portrayal of George Washington as Black or WWI soldiers as women.

"Let's learn from that absurd episode," Vance said. "The Trump team will make sure U.S. AI is bias-free ideologically and protects free speech."

Alondra Nelson, a ex-Biden science advisor at the speech, sees the "ideological bias" focus as acknowledging long-standing algorithmic issues impacting housing, loans, health, and more.

"Saying AI has ideological bias means you're spotting algorithmic bias – something we've flagged for years," said Nelson, who helped craft AI civil rights guidelines.

Yet, she doubts teamwork is possible with the bashing of equity work.

"In this divided arena, it's tough," she added. "What we call algorithmic discrimination versus their ideological bias will sadly be treated as separate issues."

This evolving debate on AI bias, DEI in tech, and government oversight could reshape how companies build and deploy artificial intelligence moving forward. Stay tuned for updates on Trump AI policy and its effects on innovation.