[robinjfisher]

I'm surprised that the guide doesn't address automated decision making about users. In my day job, we are using a scheduling platform for short-notice, short-term work e.g. social care. A job is loaded on the sytem and pushed to x users for acceptance (fastest finger first).

The system has a ratings module (not active yet) whereby clients can rate the worker and vice versa. The system then makes decisions on future job releases based on the ratings. Part of the reason for not having the module active is due to the issue of communicating to the affected users how those decisions are being made and providing them with a right of manual review.

As organisations increasingly rely on AI or ML to make decisions affecting individuals, so the need for greater transparency into those decision making processes so they can be communicated to the people concerned.