Comment by Kevin Paeth

Reporting and understanding these flaws is crucial to improving AI safety, but the reporting system is fragmented: researchers who identify flaws often do not know what or where to report, and information is rarely shared between all relevant stakeholders. The result is duplicated efforts to overcome different reporting forms and practices across private and public institutions. [...] To address these gaps, the team built FLARE-AI, an open-source, user-friendly flaw reporting system that uses conditional logic to keep required fields minimal while still capturing what recipients need for triage. It enables reporters to submit a single standardized report and optionally disseminate it to multiple developers, coordinators, and incident registries at once -- cutting down on the duplicated effort that currently discourages good-faith reporting.