Proton is asking employees to turn AI into better work—and attaching potential employment consequences to an assessment of how they do it. The company told staff on October 26 that selected employees would be re-interviewed about their AI use, according to an internal email obtained by 404 Media’s Joseph Cox.
The interviews are scheduled for November 4–11, 2026. Interviewers will document their assessments for a broader review of the organization and its talent. That review may change employees’ roles, responsibilities and continued employment, the email says. It describes a process that could lead to employment decisions, rather than announcing those decisions.
An assessment beyond the employee’s own team
The process will reach across Proton, but not every employee will be re-interviewed. Those selected will meet an interviewer outside their team and reporting line. The email describes a pool of leaders chosen for strong AI capability who have aligned on the assessment criteria.
CEO Andy Yen has already interviewed more than 40 senior leaders across Engineering and Product to begin the process, according to the email. The next stage takes that exercise beyond the senior leaders Yen interviewed, making employees’ use of AI a subject of formal review elsewhere in the company.
We need people who can translate AI’s capabilities into better products, stronger execution, and new ways of working — and who approach this change with curiosity, judgment, and urgency
Internal Proton email obtained by 404 Media
Interview assessments feed a wider review
Proton tells employees that any decisions affecting their employment will be communicated separately, with individual outcomes delivered once decisions are finalized.
Management also acknowledges that the process creates uncertainty. The email promises to distinguish what has been decided from what remains open and to communicate individual outcomes as soon as possible. Proton did not immediately respond to 404 Media’s request for comment.
Reader comments
Newest comments first. Replies stay oldest first.