By Katherine Blunt
At a moment when much of the tech industry would rather not talk politics, Google executive Jeff Dean has been a rare outspoken voice -- even on issues that hit close to home.
Dean, chief scientist at Google's DeepMind AI lab and a pioneer of the company's artificial-intelligence strategy, was among more than 30 employees at Google and OpenAI to sign an amicus brief in favor of Anthropic's lawsuit against the Defense Department this month.
In granting the AI company an injunction last week against the Pentagon's attempts to sever its contracts, Judge Rita Lin referred to, among other things, the large number of amicus briefs on the company's behalf.
The dispute, over the limits of military AI use, has large potential ramifications for Google's own relationships with the federal government. The brief Dean signed argued that the Pentagon's decision to label Anthropic a supply-chain risk would "chill open deliberation in our field about the risks and benefits of today's AI systems."
Earlier this year, Dean was one of the first and most prominent executives at a large tech company to condemn the killing of Alex Pretti by federal agents during protests in Minneapolis over immigration enforcement, calling it "absolutely shameful."
On X, where more than 400,000 people follow him, he has used his feed to boost messages criticizing the Trump administration's criminal investigation of Federal Reserve Chair Jerome Powell, celebrating diversity and immigration, and decrying the rise of measles cases as well as inequality in America.
None of these positions would have seemed remarkable a few years ago, when tech companies exhorted their employees to "bring your whole self to work" and employees responded en masse by engaging in #MeToo and Black Lives Matter demonstrations.
Since then, many big tech companies and startups have narrowed the limits of political speech by employees, saying it can create division in the workplace and distract from the mission. Some, including Google, have enacted rules on what workers can discuss on company platforms or social-media postings.
In September 2020, Coinbase Global Chief Executive Officer Brian Armstrong said the company would discourage internal discussion of societal issues or political causes not directly related to its work, arguing that excessive focus on them made the company an unwelcoming workplace for some employees. Many entrepreneurs and investors celebrated Armstrong's move, and other companies followed with similar policies.
Then Donald Trump returned to the White House, with a swath of Silicon Valley CEOs backing him or signaling support for his agenda with donations and White House appearances. His administration issued executive orders against diversity, equity and inclusion and "Woke AI" and threatened broadcasters' licenses over comedians' jokes.
"We are getting rid of woke," Trump said this past summer after signing an AI-related order. "The American people do not want woke Marxist lunacy in the AI models."
Across the tech industry, employees who previously thought little of signing their names to petitions or calling out their leaders in all-hands meetings reassessed the risk.
"Because the job market is not necessarily superrobust, people likely made the wise calculation that the juice isn't worth the squeeze in saying something, even though morally, it's very difficult for people," said Beth Steinberg, a former executive at Chime Financial who advises tech companies on human-resources strategy. "I think they are very afraid of retribution."
That new reticence -- and Dean's own reputation as one of AI's leading lights -- have made his pronouncements hard to miss.
Dean, who has a Ph.D. in computer science from the University of Washington, joined Google in 1999 after working for a computer-research lab and an online-shopping platform. He has been instrumental in guiding Google's AI ambitions, having helped develop the neural-network technology that underpins today's large language models and a long list of other advances.
Amin Vahdat, chief technologist for AI infrastructure at Google, said Dean's research over more than two decades has made him one of the most widely respected computer scientists in the world. Within the company, Vahdat said, employees see him as a kind and thoughtful colleague who has remained humble after all his accomplishments.
"It's just this stunning array of not like, nice pieces of work, but a category-defining effort, year after year after year," Vahdat said. "These papers have been cited, in some cases, literally tens of thousands of times."
No CEO of a large tech company had issued a public statement about Pretti's killing, or the wider immigration action in Minneapolis, when Dean posted on X, "Every person regardless of political affiliation should be denouncing this."
A number of Dean's colleagues sent private messages to him expressing gratitude, people familiar with the matter said. In the days that followed, OpenAI CEO Sam Altman told employees that "what's happening with ICE is going too far," while Apple's Tim Cook told employees he had discussed the need for de-escalation with President Trump.
Co-workers again looked to Dean after the Defense Department designated Anthropic a national-security threat. The company had been seeking to negotiate guarantees into its contract that the military wouldn't use its AI models for mass surveillance or autonomous weapons.
"Mass surveillance violates the Fourth Amendment and has a chilling effect on freedom of expression," Dean wrote last month on X in a post that got a million views and more than 4,000 likes. "Surveillance systems are prone to misuse for political or discriminatory purposes."
More than 100 employees within DeepMind sent him a letter in February asking him to ensure any deal with the Pentagon prevents Gemini from being used in domestic surveillance or piloting autonomous weapons. A person familiar with the matter said Dean engaged in conversation with some of the people who signed it.
At a conference hosted by Nvidia this month, Dean said he was particularly excited about the potential for AI to transform education and healthcare. He pointed an audience of several hundred people to his contributions to a project called Shaping AI, a collaboration among researchers and academics meant to foster conversation about how AI is used.
"There's obviously a whole range of things that AI can be applied to, and some are incredibly societally beneficial," he said. "Some are, you know, potentially things we maybe don't want a lot of use."
Google historically prided itself on having an open work culture that encouraged debate among employees.
That has changed in recent years as the company sought to quell opposition to some of its more-controversial business endeavors, notably a $1.2 billion contract Google and Amazon.com signed with the Israeli government in 2021.
John Palowitch, a former Google DeepMind research scientist, said he would like to see Dean use his clout among company executives, including Alphabet and Google CEO Sundar Pichai and Demis Hassabis, the DeepMind co-founder and chief executive, to affect company decision-making going forward.
"I remain hopeful he continues to speak out like this on social media and elsewhere, and even more hopeful that he leverages his relationships with Sundar and Demis to make real change," Palowitch said.
Write to Katherine Blunt at katherine.blunt@wsj.com
(END) Dow Jones Newswires
March 29, 2026 12:00 ET (16:00 GMT)
Copyright (c) 2026 Dow Jones & Company, Inc.
Comments