Internal email shows directors were warned of 'illegal' AI system and kept sharing Danish patient data

Regions Hovedstaden (the Capital Region) and Sjælland (Zealand) shared sensitive information about children with mental illness, people in cancer treatment and other patients with an American AI tool. The Danish outlet Ingeniøren revealed this in September, and it caused an outcry among the children's parents and several experts.

A new report, based on an internal email, shows that the top management was warned the AI system was "illegal" and that data sharing nevertheless continued. The email's recipients were Mette Harbo, director of CIMT, deputy director Anna Olga Aaskilde Laursen, and Kim Veber Carlsen, acting unit head responsible for the Sundhedsplatformen (the regions' electronic patient record system).

Over 100,000 patient messages containing sensitive information had been shared before the regions shut the AI trial down. Experts criticise the senior management for ignoring what they call a "clear warning light."

Earlier coverage from the same outlet adds context. A large investigation into the "illegal" AI trial has been launched, and one source says those involved "must be held to account." Parents of sick children were told nothing about the trial, and one mother says she fears what has been shared about her child. A senior physician has also voiced concern for Danish health data and wants a new region to move away from American tech.

The story matters because it points to a failure at the top of the organisation rather than a mistake at lower levels. It also raises questions about how Danish health data is handled when sent to US-based AI services.