Why ChatGPT is not proof by itself
Enterprise ChatGPT gives you a walled history for that workspace. It does not see Copilot in Word, Slack AI in a client channel, or the personal ChatGPT account used after hours on the same facts.
Shadow ChatGPT is the event firms discover during a scare. If it is not on the same trail as sanctioned use, the export you hand a client is a fiction of completeness.
Proof on the Taigris row
Taigris does not replace ChatGPT. It puts that surface on the proof trail (source, user, system, policy, hash) with the rest of the toolbox.
- System distinguished: ChatGPT vs other frontier clients vs Copilot.
- Policy distinguishing sanctioned workspace use from shadow visits.
- Attribution of user and time where the environment can observe it.
- The same hash chain as every other AI event.
Customer-owned environment
Public ChatGPT on unmanaged devices is a detection-and-policy problem, not a magic API. Taigris puts those events on the trail when they are in scope It does not claim omniscience.
Related