Proton has apologized to its users after growing anger over its AI plans. The Swiss company runs Proton Mail, Proton VPN, and other privacy-focused services. Some customers have threatened to cancel their subscriptions.
The criticism followed reports by 404 Media. Those reports said CEO Andy Yen pushed staff to use AI tools more aggressively. Some engineers reportedly raised worries about security and the way the company builds its software.
An internal email added fuel to the fire. It revealed plans to re-interview workers about how they use AI. The answers could affect their roles and their jobs. Proton posted its reply on October 30, 2026. The company admitted it had not met the standards its community expected.
Proton has also promised more openness about its internal AI habits, code reviews, and security rules. Reports about its workplace culture pushed the company to make those promises.
Proton denies tracking workers and promises more openness
Proton denied putting tracking software on laptops to count AI code from each worker. The company said the claim came from a misunderstanding. It blamed confusion over how its mobile device management software works. That kind of software helps companies manage work phones and computers.
However, the October 29 investigation by 404 Media said something slightly different. It reported that Proton installed software on staff computers to track how much AI code the company produced. The two claims are not exactly the same. Watching total company output does not prove that the company tracked each worker.
Proton also rejected the idea that AI would weaken its security. The company said its AI use would never come at the expense of safety. It added more layers of security reviews over the past year. Those reviews aim to catch weak spots that AI-assisted coding could create.
According to Proton, the public can still inspect its open source code. The company also promised to publish how it reviews code and checks security. It will add clearer rules on where staff can and cannot use AI.
Proton made one more pledge. It plans to release some of its internal AI tools as open source projects. These include stronger versions of Lumo, its privacy-focused AI assistant. Proton says those versions have only been available inside the company so far.
Staff re-interviews cancelled after worker and customer concerns
The story gained attention on October 26. That day, 404 Media reported on the internal email about re-interviews. The email called the assessments consequential. It said they could change a worker’s duties or decide whether the person keeps a job.
The report pointed to the pressure on staff to show they can work with AI tools. It also showed a clash between managers and engineers. They disagreed on how much time new code needs for testing before release.
On October 29, 404 Media published a second investigation. It drew on Slack chats, internal recordings, and interviews with eight current employees and one former employee. The report described one exchange in detail.
In that exchange, an engineer asked for extra penetration testing. This means trying to break into software to find weak spots before release. Yen reportedly rejected the request. According to the report, he called that way of thinking the mindset of losers.
The backlash led Proton to cancel the planned re-interviews on October 30. The company later said it was not re-interviewing staff about AI use. It also said no one would lose a job because of such interviews. You can read more in 404 Media’s report on the cancellation.
The company’s statement followed four days of growing attention from workers and customers.
Chief technology officer Bart Butler also explained the company’s changing approach to making software. He shared an AI manifesto on November 1. He argued that engineers must now guide and supervise AI models. Writing code by hand should no longer be their main focus.
Some customers still plan to leave Proton
The apology did not settle every concern. Some users have publicly said they will cancel their subscriptions. Others plan to skip their renewals. Some people want Yen to step down. They argue the row points to deeper problems with the company’s leadership and workplace culture.
This matters because Proton sells privacy tools. Users rely on them to keep messages and online activity safe. That is why worries about code testing drew such strong reactions. No verified public figure shows how many customers have actually cancelled. Nothing confirms how much money Proton may have lost either.
Proton has previously highlighted its commitment to protecting user privacy, even when facing legal requests. The company says 458 legal orders since 2019 yielded no user data, reinforcing its position on user privacy and data protection.
Reactions also differ inside the technology community. Some industry figures understand why software engineers feel pressure to adapt to AI-assisted development. Critics disagree. They say companies that protect privacy and security must keep strong human oversight and strict testing.
Proton now faces a bigger job than explaining its AI plans. It must show that new ideas and strong security can work together. Its reputation rests on the security standards and user trust it has built.
