Generative AI at Work / Unit 6 GAI 145 worked example, 60 to 90 seconds
Reporting an Incident at Work
The idea
Because NIST notes there are no formal external channels for AI incidents yet, workplace reporting runs through the organization's own incident process, and it follows four steps.
In practice
Step one: stop the harm first: correct the output, pause the use, or warn the people affected. Step two: report through your organization's normal incident route, saying that an AI system was involved and how. Step three: preserve the evidence: the prompt, the output, the sources, and the AI-use log entry. Step four: once the facts are known, feed the lesson back into prompts, review levels or the feature gate.
Where people slip
AI involvement is not a separate kind of incident and not an excuse. The same accountability applies as for any other error; the record of AI's part makes the fix possible.
Your place on the trail
Every lesson rests on earlier ones and opens later ones. Tap any to jump.
Builds on
Unlocks
- Nothing yet depends on this.
Something wrong in this lesson? Flag it · Sources for Generative AI at Work