Privacy-focused provider Proton is using machine learning models to stop cybercriminals from exploiting its email service. The Switzerland-based organisation's anti-abuse team builds systems to detect automated mass sign-ups, bot-driven clusters and suspicious account creation patterns. Chief operating officer Raphael Auphan said end-to-end encryption means the company cannot access message contents or geolocate users. Proton relies on these account-level defences to block malicious actors before they can run their operations.

Background: Automated Decision-Making in Policing

← Back to the news feed