Researcher Paid $300 for Auth Bypass on Major AI Provider

A researcher bypassed authentication and sandbox protections on a major AI provider's paid models and internal API, and received only $300. The author argues the payout misprices the severity of the finding and creates bad incentives for responsible disclosure.

• The bypass exposed paid models and an internal API without authentication
• The payout was $300, less than a typical monthly phone bill
• Author: AI firms' bounty tables underestimate real blast radius
• Low payouts push researchers to sell findings instead of disclosing them