Like other regulation-based labeling around choking hazards, consumers need to know where there is machine learning in products and services and how (if at all) it interacts with their use.

We conceive of four main categories for products and services to identify how (if at all) they leverage machine learning:

  1. Uses ML
    These systems may leverage machine learning models to deliver functionality, but they do not contribute to models or collect and share data
  2. Collects Data
    Like the many systems covered by GDPR, these systems collect user/environment data and may transmit it and store or process that information
  3. Trains Onboard ML
    These systems may use data from the user/environment to refine a ML model, but that information stays in a closed system and is not shared or aggregated.
  4. No Data Collection or ML