OpenAI unveils privacy-first misuse monitoring to rival Anthropic's policy

As AI models grow more powerful, so does the risk they’ll be misused — and AI companies must balance respecting enterprise customers’ privacy against monitoring for abuse. OpenAI announced a privacy-centric safety approach it calls Private Safety Processing, an automated system previewing to select customers that watches for potential misuse while retaining none of the customer’s data.

The move runs counter to Anthropic’s data-retention policy, announced in July, which lets the AI lab keep user session data for 30 days for “covered models” — a category that includes all Mythos-class models and future models with similar capabilities. Anthropic designed that policy for safety purposes, to let it sift and analyze potential impropriety, but it has concerned some enterprises handling sensitive data who don’t want it retained or inspected by the lab. Anthropic says any human review of that data happens only “through a controlled access path” involving “a small set of approved reviewers,” with every review session “recorded in a tamper-proof log that reviewers cannot suppress or modify.” Anthropic otherwise largely follows Zero Data Retention (ZDR) policy, except for covered models like Fable.

OpenAI, like most AI companies, already offers customers Zero Data Retention, under which agents within its API monitor for abuse per session without retaining customer data. Private Safety Processing widens that scope: an agent assesses inputs and outputs across multiple conversations, not just one, to catch bad actors — such as someone building malware for a cyberattack — who spread requests across sessions to avoid detection, a spokesperson told TechCrunch. If the system is triggered, it may send OpenAI a “narrowly defined signal” about specific activity, which OpenAI uses to decide whether enforcement is necessary; if so, it reaches out to the customer, who may then choose to share data at their discretion.

The privacy competition comes amid tense corporate rivalry between the two companies, both racing toward IPOs. A recent report showed OpenAI’s second-quarter growth slower than Anthropic’s, whose annualized revenue run rate is now reportedly $65 billion; Anthropic investors have said it could IPO at $2 trillion.

Why it matters

The dueling privacy approaches show AI labs treating data-handling policy itself as a competitive lever for winning enterprise customers, not just a compliance requirement — a dynamic likely to intensify as both companies head toward public listings.