Machine Learning

The technology

AI and machine learning are changing the way we can tackle some of the world’s most challenging issues. Kindred Futures have set out to explore how this technology could be applied to better understand and most importantly, prevent problem gambling and addiction before it even begins.

The Challenge  

Kindred’s long-term ambition is to reach zero revenue derived from harmful gambling by 2023, and the use of the latest technologies will be key in tackling the issue.

In 2020, Kindred initiated a collaboration with Zafty and ASSISSA. The project focused on supporting the development of an app that could be used by all gamblers, operating on different platforms, to monitor and control their gambling. Specifically, the app would be able to identify problem behaviours before they develop and provide tools to control that behaviour. At the same time, the app would be able to gain hugely valuable insight into the factors involved in the development of these early-stage behaviours.

The partners

Zafty is a leading company in mobile mental health assessment, using proprietary machine-learning software to identify unique changes in user behaviour associated with mental health issues.

ASSISSA Consultancy Europe provides training for the gaming industry. This includes various communication training courses, E-learning, value analysis, policy plan development, marketing and interventions in the form of consultancy.

Together, Zafty and ASSISSA will initially launch this free to download app on the Google play store.

The outcome

On the surface, the app is essentially a time tracking app that monitors the users' time spent on gambling-related activities across all operators. The app timestamps recognised gambling products that are open and front-of-screen to measure their use.

The clever bit is going on in the background. Zafty’s unique machine learning algorithms take this activity and learn each user’s normal behaviour. Once this baseline has been established, the system can monitor changes in activity, sometimes very subtle, that might indicate early signs of problem behaviours. For example, a player regularly bets for two hours on a Saturday morning but then, one month in, starts to play at midnight on a Tuesday.

Based on the level of change in the user’s behaviour, the app will then recommend tailored features and support that will help the user bring their gambling behaviour back to a healthy level.

The screen-time tracker application will be finished in 2021. An initial launch on Android will take place in the UK and France. IOS launch will follow thereafter, with further development and partnership with other operators.