Proton CEO controversy: what Andy Yen said, and should you leave?

The reporting draws on nine current and former Proton employees, plus Slack chats, meeting recordings and an AI manifesto written by the chief technology officer. The most damaging moments came from Yen himself.

• On security testing: in August, an employee asked for a few days to run penetration tests, where testers try to break new code before it goes live. Yen rejected that in blunt, profane terms, adding that the team could be both fast and secure.
• On writing code: he told engineers to stop crafting code lovingly by hand and lean on AI instead.
• On hiring: he wrote that this year would be Proton’s final year of hiring engineers.
• On user complaints: after people on Reddit pointed out that Proton’s new post-quantum encryption, announced in May, left existing messages unprotected, Yen’s message to staff was that how things look matters more than “truth”.

One employee quoted in the investigation described the chief executive as being in “full AI psychosis” for some time. Staff said they felt unable to raise concerns for fear of being berated or losing their jobs.

A lot. In a meeting recording obtained by 404 Media, chief technology officer Bart Butler said AI wrote 90% of Proton’s code in October. Employees said much of it was produced with Anthropic’s Claude.

The change has been fast. AI use was originally banned inside the company. Proton has since installed software on staff computers that measures how much of their code comes from AI. An internal document also says Proton is behind its target for annual recurring revenue this year, and it blames “insufficient velocity”, meaning the company thinks it is moving too slowly.

That context explains the pressure, but it also explains the alarm. AI-written code isn’t automatically unsafe. The worry is speed without checks, because security testing is exactly the step that catches mistakes before attackers do.

Before the investigation, 404 Media revealed an internal email saying Proton would “re-interview” selected staff about how they use AI. The email called the interviews consequential and said a wider review could lead to changes in roles and in continued employment.

Yen started by interviewing more than 40 senior leaders in engineering and product. Interviewers were to come from outside each employee’s team, drawn from leaders said to show strong AI ability, with sessions planned for November 4 to 11.

Proton pushed back on Reddit, saying it wasn’t using AI as cover for layoffs, that it faces a talent shortage, is still hiring, and wants to find who needs more support. Then, as the backlash grew, a later report said Proton will no longer hold the re-interviews at all. The company hasn’t detailed what replaces them.

A Proton spokesperson didn’t dispute the reporting. The company’s response conceded that in its rush to move fast it had sometimes been too blunt with its staff, or too slow to explain itself.

Customers were less forgiving. On Reddit, users posted that they were leaving, with one arguing Proton had been the opposite of a big tech company and that customers, not the CEO, decide what counts as quality. On Proton’s own user forum, a post calling for Yen to resign gathered 831 likes and dozens of comments.

Yen has argued before that privacy companies must build their own AI, or they hand the future to firms that don’t have users’ interests at heart. Proton’s chatbot, Lumo, is built around that idea: the company says it doesn’t log conversations or train on user data.

Nothing in the reporting says Proton’s encryption has been broken or that any user data has leaked. The controversy is about how the company builds and tests its software, not about a known security flaw. Your existing emails stay encrypted the way they were before.

The real question is trust over time. People pay Proton because they believe it is more careful than the giants. A CEO waving away security testing cuts against that, even if nothing has gone wrong yet. If you’re a subscriber, here is a sensible approach:

• Don’t panic-cancel. There’s no reported breach, and moving your email is a big job that can cost you messages if rushed.
• Turn on two-factor authentication if you haven’t, so your account is protected whatever happens behind the scenes.
• Watch for Proton’s next moves. Independent security audits, clear testing policies and how it handles the next bug report will show whether the culture has changed.
• Know your exit. If you decide to leave, look up Proton’s export tools first so you can take your mail and contacts with you.