The company treats it as a misalignment case and separates it from the Hugging Face breach. But the confirmation was not volunteered: executives knew weeks earlier and the statement followed the Reuters story.
Detection tools produce probability, not certainty. Four families of methods, the three limits of watermarking, the cost of a false positive and the extra difficulty Turkish adds — a practical framework for how an organisation should use these tools.
Protect Democracy has sued four federal agencies to force the release of the framework used to review frontier models before launch. The framework is not classified, yet it still is not being shared.
The company states the reason plainly: people dislike finding out later that a profile was AI. Those who use the label are not penalised.
According to the Washington Post, members of Congress and their aides use AI tools with little oversight for everything from writing speeches to sorting constituent mail.