Navigation

Increasing expertise in healthcare thanks to artificial intelligence

Smartwatch
Image: Colourbox

FAU is leading a collaborative research project about the interaction between humans and AI to promote health

How should an intelligent assistance system be designed so that as many people as possible use it to improve their health on a daily basis? The collaborative project ‘Erweiterte Gesundheitsintelligenz für persönliche Verhaltensstrategien im Alltag’ (Eghi) (Advanced intelligence in healthcare for personal behavioural strategies in everyday life) which has now started under the leadership of FAU is focusing on this question. The Federal Ministry of Education and Research (BMBF) is providing around 1.8 million euros in funding for this project over a period of three years as part of the ‘Adaptive Technologien für die Gesellschaft – Intelligentes Zusammenwirken von Mensch und Künstlicher Intelligenz’ (Adaptive technologies for society – intelligent interaction between humans and artificial intelligence) funding priority.

‘Even though people’s interest in and commitment to improving their health is increasing on the whole, it’s often difficult for individuals to incorporate the relevant measures into their daily lives,’ says Prof. Dr. Oliver Amft, who is the coordinator of the collaborative project. He is holder of the Chair of Digital Health at FAU and emphasises: ‘Existing services are often only used by people who already take good care of their health. So-called wearables such as fitness trackers and smartwatches are already very popular in this group.’

A learning assistance system

In collaboration with the German Research Centre for Artificial Intelligence (DFKI), the University of Duisburg-Essen, BODYMED, and Interactive Wear AG, FAU hopes to develop a self-learning assistance system based on artificial intelligence (AI) that will help people to develop healthier habits in their everyday lives. ‘We are focusing on personalised, situative recommendations for behaviour that are directly linked to the experiences people have and that are easy to put into practice,’ explains Professor Amft. These recommendations could be anything from reducing health risks to incorporating more exercise into your daily routine.’

The researcher gives the following example: ‘A user travels by tram to an appointment. During the journey, the intelligent assistance system Eghi detects that the weather is good and that there’s enough time before the appointment. The system has also often detected that the user likes to go for walks if the weather is nice. Eghi therefore asks the user if she would like to get off the tram earlier and walk the rest of the way so she can get some exercise.’

Whereas such a recommendation would be obvious to a person in this situation, artificial intelligence needs to model it first. ‘This is a complex problem and one we hope to find a solution to in this project,’ says the project coordinator. ‘The processes for generating personalised situative interaction and suggestions for which actions to take do not yet exist.’

In addition, methods in AI that actively incorporate users into decision-making processes in an understandable way and that support users in forming personal behavioural strategies as a form of enhanced intelligence are not yet available either.

Interdisciplinary approach to research

The aim of the Eghi project is to apply the concept of enhanced intelligence to support healthy habits in users’ daily routines and help them to form personal strategies for behaviour. The project team is taking an interdisciplinary approach to its research and is linking methods from artificial intelligence with behavioural modelling methods and concepts for interaction between humans and technology.

The first step involves creating a type of joint collection of experiences using sensor-based observation of users’ activity and their behavioural patterns and creating a joint communication platform between humans and AI.

The next step is to consolidate the user data using AI in such a way that the system can derive and make personalised recommendations for actions the user could take.

‘These recommendations are easy to implement and to understand as they are linked to things the user has experienced,’ says the project coordinator. In a well-established relationship between humans and AI, a short buzz from the user’s smartphone or a symbol flashing on a wristband could remind the user to prepare a healthy meal for dinner.

‘The aim of the notification is to give users new opportunities to become more health conscious. We are developing an intelligent assistance system for healthy eating, exercise as part of users’ daily routines or for supporting older users, but not a monitoring system,’ emphasises Prof. Dr. Oliver Amft.

Further information

Prof. Dr. Oliver Amft
Chair of Digital Health
Phone: +49 9131 85 23601
oliver.amft@fau.de