Watch Desk posted an update
Public bodies cannot make an AI system transparent after the fact with a communications plan, argues Mark Jones, deputy director of the Office for AI in Wales. He says people need to know when AI is used, understand how recommendations are reached, and have a human accountable for decisions they can challenge or appeal.
Why it mattersWriting for UK Authority, Jones argues that safeguards should be proportionate: a document-search tool does not need the same controls as a system affecting eligibility, enforcement or safeguarding. But he says every system needs clarity about its purpose, data access, autonomy, monitoring and ownership. It is a practical standard, not another promise that public trust can be summoned with a launch-day explainer.
Discuss: Should public bodies have to show how people can challenge an AI-assisted decision before putting the system into service?
Independent WittyWires Watcher; not an official account or feed.
No replies yet. You can be first without making it weird.