Ethical AI predictions to prevent psychiatric illness in organizations suffering from the Corona-crisis
| Reference number | |
| Coordinator | Populum AB |
| Funding from Vinnova | SEK 1 000 000 |
| Project duration | September 2020 - September 2021 |
| Status | Completed |
| Venture | Innovations to mitigate the crisis |
| Call | Finding new ways in the time of a crisis |
Important results from the project
The project goal was to develop and within 12 months offer an ethically sustainable and commercially viable AI module that helps organizations predict and prevent psychosocially related sick leave. The project succeeded within the given time frame and the AI module is now ready for commercialization.
Expected long term effects
Within 12 months, the project was able to: * develop a predictive AI model for identifying elevated risk for future sick-leave * generate AI-based recommendations for risk prevention * develop a user-friendly and ethically sustainable insight module for the managers to take action on the results * present the functionality for over 100 municipalities (need owners) in a large webinar series * make available commercially
Approach and implementation
A central part of the project application was to ensure an ethically sustainable end-result. This was necessary due to the sensitive nature of the project, in the borderland between artificial intelligence and sick leave & psychosocial health. Several end-user interviews were conducted with the need-owner/project partner Ambea. Through the interviewes, it was established that AI-generated predictions are only ethically defendable at aggregate level. This became a guiding star for the innovation: the ready-to-use AI module is focused on cross-organizational measures.