Mindway AI raises concerns about a detection model
•
21 Sep 2026 •
Reading time: 3 minutes
Can a model based on two years’ worth of player data really identify people at risk of gambling-related harm? Mindway AI believes that the tool developed by researchers at the University of Amsterdam primarily measures the likelihood of a player voluntarily self-excluding, without necessarily identifying gambling problems.
A model that raises a key question
To build this tool, they drew on two years’ worth of data from casinos. The information taken into account includes, in particular, the moment when a player decides to self-exclude from an establishment.
It is precisely on this point that Mindway AI has reservations. According to the company, predicting that a person is likely to request self-exclusion does not necessarily mean predicting that they will develop gambling-related harm.
Self-exclusion is not a diagnosis
Kim Mouridsen, founder of Mindway AI and professor of neuroscience at Aarhus University, points out that self-exclusion was originally designed as a way for a player to stop gambling quickly.
‘Self-exclusion is intended as a safety measure, not as a diagnostic indicator.’
According to Mouridsen, research also shows that the majority of people who self-exclude do not have gambling problems. Conversely, around 80 per cent of those in need of help would not request self-exclusion.
Annika Lindberg, a psychologist and member of Mindway AI’s expert panel, shares these reservations. She points out that a person suffering from serious gambling problems may continue to gamble without ever requesting exclusion. There are several possible reasons for this, including a poor perception of their own behaviour or difficulty in stopping.
‘Self-exclusion is undoubtedly linked to gambling problems, and many of the behavioural characteristics used are highly relevant for assessing risk. The question is what we can actually deduce when a person is flagged in this context.’
What behaviours should be monitored?
The company believes that more accurate detection should be based on several behavioural characteristics simultaneously.
Rasmus Kjaergaard, CEO of Mindway AI, cites in particular the tendency to try to recoup losses, a gradual increase in the time spent gambling, and certain impulsive behaviours during transactions. According to him, these signs may appear before a request for self-exclusion is made.
Mindway AI states that it has developed its own models to analyse different risk factors separately. The company also emphasises the role of personal circumstances, the environment, neurodiversity and emotional wellbeing in a player’s behaviour.
The algorithms are already in use by casinos
According to the company, the distinctive feature of the model developed by the University of Amsterdam lies primarily in its public accessibility. Researchers, regulatory authorities and other stakeholders can therefore examine how it works and make use of it.
This openness may facilitate cooperation and the comparison of methods. Mindway AI considers, however, that it does not resolve the fundamental question of what the model actually measures.