Proton Cancels Employee AI-Use Re-Interviews as Security-Check Dispute Surfaces

Proton has canceled a plan to re-interview employees to assess how they use AI. Citing a recording of the meeting, 404 Media reported that CEO Andy Yen announced the reversal at an all-hands meeting on October 30 local time. At the center of the controversy is how the company handles employee evaluation and security review as it pushes to expand AI use.

A few days for a security check: why it became a problem

The dispute began over how much time engineers get to check security before new code ships. According to 404 Media, an employee asked in late August 2026 for a few days to run a penetration test, an exercise in which testers attempt real attacks to find weak points in security. 404 Media reported that Yen criticized the request and pressed for greater use of AI in development.

404 Media said it spoke with eight current employees and one former employee. It also said it reviewed internal messages and meeting recordings. The reporting focuses on how management responded to an employee who asked for checks before a product shipped.

The report also includes an internal remark from the chief technology officer (CTO) that AI wrote 90% of the code. In effect, the push to use more AI in day-to-day engineering came into conflict with the time needed for security checks.

What the canceled re-interviews were meant to measure

An internal email released by Proton laid out a plan for re-interviews assessing how employees use AI. That plan was withdrawn after Yen announced the cancellation at the all-hands meeting. Because the company’s push to use more AI was about to feed into a process for evaluating individual employees, the issue went beyond which development tools people choose.

Proton said it is not using AI as a justification for cutting jobs and that it is still hiring. The company also said it would weigh privacy, security and accuracy when assessing the use of AI tools. It added that it would provide the support employees need to use AI. Yen acknowledged that the company’s internal communication had been flawed and said no one would lose their job to AI.

What users should actually watch

For users, what matters is less how much AI is used than what security review a product goes through before it ships. The point to watch is how the privacy, security and accuracy Proton has pledged are reflected in its actual development process.

It was the re-interview plan that was canceled. One employee told 404 Media that the AI-use metrics the re-interviews were meant to measure will be folded into regular performance reviews. The thing to watch going forward is what procedures the company puts in place to deliver the employee support and security considerations it promised. The conflict over pressure to expand AI use, the time required for checks and how employees are evaluated underscores the need to examine changes in tools alongside changes in how people work.