OpenAI previews private safety processing, building on 0 data retention
OpenAI is previewing Private Safety Processing, a new system that helps catch people misusing AI, without retaining users' data.
It builds on its Zero Data Retention policy, so users' information stays with them unless they choose to share more.
Automated agents look for sketchy behavior, like someone trying to engineer malware for a cyberattack, and only send OpenAI very limited alerts if something seems off.
Anthropic retains user data 30 days
This move might win over companies worried about privacy, especially since competitor Anthropic keeps user data for 30 days for covered models, which some find risky.
With both companies racing toward IPOs and battling for the top spot in AI, OpenAI's privacy-focused approach could give it an edge with users who care about keeping their data safe.