ChatGPT Remembers Your Disability — And That's a Privacy Problem
Memory features spare disabled users from repeating themselves — but leak private details.
Researchers interviewed 12 adults with disabilities in the United States who use LLM-based assistants like ChatGPT, Claude, and Gemini, asking when, how, and why they disclose disability to these systems — and how that compares with disclosing to people. Using contextual integrity as an analytic lens, the study found that participants disclosed by need rather than by name, translating disability into task-scoped instructions. The same disclosure was judged against two recipients — a non-judging interlocutor and a data-holding company — producing opposite norms. Memory features relieved the burden of repeated disclosure while also letting disability information drift into contexts where it did not belong. Participants did extensive boundary work to restore context, and they wanted control over scope, provenance, retention, and access rather than per-utterance toggles. The paper is under review.
- Most disabled users never say "I'm disabled" — they ask for tasks, so the AI helps without ever labeling them.
- Memory saves them from repeating themselves, but lets private details drift into unrelated conversations.
- Users want control over what's remembered, how long it stays, and who can see it — not just a delete button.
Why It Matters
If your AI remembers you, it can also expose you — everyone with sensitive chats needs better controls.