Apple asks 200,000 people to grade a health chatbot
A study invites people to ask an AI their health questions, with a promise that the answers will not train a model.


Picture made for ABUREAU.
Would you ask a chatbot about a worry you have not told your doctor? Apple wants to learn how people do exactly that. Its new study is open to up to 200,000 adults aged 21 and over in the United States who own a recent iPhone.
Taking part means asking everyday health questions in a study app and rating the answers in surveys. Sharing personal health records is a choice, not a condition. Apple says people control whether they share any HealthKit metrics, such as resting heart rate, activity, sleep or lab results.
The notable line is a limit on use. Apple states that the "study data will not be used to train AI models". The announcement does not say how long the data is kept or how long the study runs.
For anyone who owns or uses data, this is a clear pattern. Personal data comes with a stated limit on what it may be used for, and the participant picks what to share. A limit on training is a term that can be written down and checked. Where retention is not stated, that is the next question a careful owner or buyer asks.


