i
News
News · 2026-11-05

Google employees say they cannot speak openly about AI risks

Dataist Dataist

Google employees say they are being prevented from telling the public the truth about AI

Cover: Google employees say they cannot speak openly about AI risks

Employees at Google DeepMind’s London lab are afraid to speak publicly about AI’s problems and potential harms. According to The Washington Post, a broad non-disparagement agreement is getting in the way. Dozens of workers discussed its effects at a meeting organized by colleagues. Some compared the document to the surveillance state in George Orwell’s novel 1984. Employees consider the agreement’s terms excessive, even if the company could not enforce them in court.

DeepMind employees accuse Google of banning talk about AI risks

Why employees are afraid to speak

Non-disparagement agreements prohibit workers from disclosing information a company might consider negative or critical. According to The Washington Post, DeepMind includes such a clause in documents many employees signed when they joined the company. It worries those who would like to inform the public about the technology’s development and its associated risks.

One DeepMind employee, who asked to remain anonymous, told the publication that AI is changing rapidly and could have serious consequences. The public, the employee said, wants and deserves to know what is happening, and information about harms and potential threats often becomes known through accounts like these.

Two employees told The Washington Post about a discussion group their colleagues organized to talk about the non-disparagement clause. Dozens of people took part; some openly criticized the agreement and compared it to the surveillance state in George Orwell’s 1984.

Workers are unsure whether Google’s leadership will enforce the clause or whether the company can legally do so. But, one employee said, even if the agreement is legally unenforceable, the attempt to restrict speech is intended to intimidate people.

Dataist’s take: The agreement may affect employees’ willingness to speak about AI risks before any dispute over its legal force arises. The article does not reveal what specific information the workers would like to disclose.

Daily AI news

Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.

Only what matters — every day

Follow on X