Public Employment Service Austria (AMS)
AMAS job-seeker prospects scoring model
Austria's job agency built a job-seeker scoring model, lost it to a data-protection ban in 2020, and won in court five years later too late to use it
Problem
AMS had limited money for training and support and wanted to spend it on job seekers who would benefit most. Advisors made that call case by case.
Approach
AMS commissioned the outside firm Synthesis Forschung to build statistical models on past employment records. The tool gave each job seeker a score and sorted them into three groups: good prospects, medium prospects who would get the most support, and low prospects. Advisors were meant to keep the final say. A nationwide launch was planned for 2020 to early 2021.
Results
Takeaway
The ban came before launch and forced AMS to delete the data, so the model never went live. A court win in 2025 did not bring the project back.
Categories
- Size
- Large enterprise
- Industry
- Public sector
- Business function
- Production and quality
- Innovation type
- Process innovation
- AI technology
- Predictive analytics and ML
- Adoption
- Co-developed with a partner
- Evidence
- Company's own claim
Sources
Source summaries
APA reports that the Federal Administrative Court annulled the 2020 data-protection decision against AMAS more than five years later. It describes the three-group scoring and says advisors were to decide. AMS head Kopf said the 2020 ban wiped out work worth about €2.5 million and that the project will not be revived.
AlgorithmWatch described the planned scoring system, its three groups and the 96 models behind it, and put the outside analysis cost at €240,000. The AMS documentation showed negative weights for women, disabled people and people over 30. Austria's data-protection authority declared the system illegal in August 2020.
Full details are in the original sources.