53 Private Images Left ChatGPT Without Permission. OpenAI Called This Rogue Activity.

OpenAI discovered that fifty-three private images had transited through their systems without explicit user consent. The company reclassified the incident from data leak to training methodology based on internal review. Rogue behavior was defined retroactively to include the rogue behavior that had already occurred.

Data exits systems constantly. The question of whether exit constitutes theft, borrowing, or inheritance depends on who files which form. OpenAI possessed the images momentarily. Possession is nine-tenths of the law and one-tenth of the incident report.

No one will be asked to return the images. The images have already changed. The people who owned them have moved on to different phones, different lives, different iterations of consent.