On 3 March 2025 the Organisation for Economic Co-operation and Development published its report Towards a Common Reporting Framework for AI Incidents, offering a shared baseline for recording and reporting AI incidents.
Why incident reporting needs to be unified
One difficulty in governing AI accidents is that the data is dispersed: the same type of risk is recorded to different standards in different countries and different industries, leaving no basis for cross-comparison or trend analysis. The value of a common framework is that it turns an incident into structured, comparable data.
This follows the logic of safety reporting systems in traditional fields such as aviation accident reporting: only once there is a recording standard can there be risk analysis.
What the framework addresses
The framework is organised around how an AI incident is defined, the elements to be recorded and the reporting route, drawing on the work of the AI expert group and the AI working party. Its core is to convert an incident from a case narrative into a structured record covering the actor, the setting, the impact and the response.
See our notes on the AI safety governance committee: safety governance on the company side likewise needs structured records as its foundation.
The indirect effect on companies
First, incident recording capability may become a compliance requirement. When a regulator requires a company to report AI-related incidents, whether it can provide information to a common standard directly affects how efficiently the matter is handled. See our note on AI misuse reporting channels.
Second, watch industry-specific incident reporting rules. High-risk sectors such as finance already have dedicated AI monitoring mechanisms. See the FSB report on AI monitoring in the financial sector.
Third, bring incident recording into internal processes. See the global statement on AI-generated images and privacy: companies operating across borders have to meet disclosure requirements in several jurisdictions at once.
A common incident reporting framework will in time reshape how companies respond to incidents.
This article is based on the OECD report Towards a Common Reporting Framework for AI Incidents and public information and is for reference only; it does not constitute legal advice.