I'd wager it's more due to decision making culpability.
Now, for the next "teen committed suicide after talking to AI therapist" headline, Claude can deflect the blame to the parents for not supervising their child enough.
If you have an ineffective system for preventing a thing from happening, but one where the "victim" has to override the system, the victim is at fault.
Your honor, the user manual in the box did say "don't put children in the microwave."
Because in most of those API, even many implementations of the Responses API, you lose a lot of control of where your data is going. e.g. an Agent or the Responses API may automatically invoke a tool call that leaks your data to an external service on the internet, without having an option to intervene.
If you want to have control over your data, you have to have control over your harness.
Photographing bikes or electric scooters does nothing without enforcement through the city. It allows tracking down the person who placed it wrong, but in the absence of fines through the city, the rental companies surely won't start fining their customers and drive them away.
The Lime app (bikes and scooters) refuses to close down a trip if you enter an area where parking is forbidden or if the picture shows that the vehicle is blocking the sidewalk or pedestrian paths.
It will also 100% forward you a fine if they get one and their records show you were the last user who misplaced it.
In a world were there is a significant amount of people that do things just to take photos of themselves doing things rather than experiencing the thing, this may be the lesser evil.
The distiction they are trying to make is: "One of our employees or the model was able to verbatim read the chats when they were actively tackling the problem" vs "The chat of someone working on the problem may have ended up in the training set of the model".
Regulation != censorship. (Apart from that, Mistral has Shieldstral[0] for adding policy enforcement).
With it's general approach of Open Weights, and being able to be deployed on-premise (/private/public cloud), they are a viable solution for your typical European enterprise, that has to comply with traditional (non-AI) regulation. With the NIS2 directive the amount of those companies is also significantly expanding.
We[1] are in the same market as Mistral, and among our customers, the go-to-solution of MS Copilot is typically performing badly, and the typical SaaS solutions are not even given a consideration, which is why they are reaching for on-premise-first solutions.
reply