Report incident 🔥 | Improve page 💁 | Access database 🔢
The so-called "AMS algorithm" was developed in 2018 by Austria's Public Employment Service (Arbeitsmarktservice or "AMS") to predict a job seeker’s employment prospects and allocate appropriate forms of support to them.
The system works by automatically classifying job seekers and calculating individual 'IC' scores based on their gender, age, citizenship, education, health, care obligation and work experience, amongst other factors, to determine their relative employability.
It then assigns them to one of three possible prospective employability groups - A (High), B (Medium), or C (Low) - with those in the lowest category denied access to expensive training courses. Job seeker scores can be petitioned against and overriden by human case workers.
The system was supposed to go into regular operation on January 1, 2021, but the start was delayed due to the COVID-19 pandemic.
Website: AMS 🔗
Operator: Public Employment Service (AMS)
Developer: Synthesis Forschung; Public Employment Service (AMS)
Country: Austria
Sector: Govt - employment
Purpose: Assess employability
Type: Prediction algorithm
Technique: Linear regression
Johannes Kopf (2019). Ein kritischer blick auf die AMS kritiker
Synthesis Forschung (2018). Das AMS-Arbeitsmarktchancen-Modell (pdf)
Several transparency and accountability limitations have been associated with Austria's AMS algorithm:
Data disclosure. Only 2 out of 96 claimed statistical models used in the algorithm have been made public, with one released only upon request. The error rates and full details of the 96 model variants are largely unknown.
Public information. Obtaining information required extensive correspondence and multiple inquiries, rather than proactive disclosure by AMS. Critics argue that AMS, as a public entity, should voluntarily provide verifiable facts and data to enable proper scientific analysis and democratic discourse.
Impact assessment. There was no algorithmic impact assessment conducted before the system's implementation, which could have helped identify potential issues. The system was not adequately piloted or externally evaluated before being applied to structurally disadvantaged groups.
Algorithmic decision-making. The decision-making process of the algorithm is not easily understandable or explainable to the general public and policymakers. There is a lack of transparency about how the algorithm affects practical counseling and allocation of subsidies.
Complaints and appeals. Complaint options for affected individuals have not been clearly communicated or published.
Austria’s AMS algorithm has been criticised for mainfesting several forms of discrimination and for reinforcing structural prejudices and stereotypes, failing to capture or analyse soft skills and motivations, violating privacy, and exhibiting poor transparency.
August 2020. Privacy watchdog halts Austria job seeker prediction system
October 2019. Scientists criticise AMS algorithm transparency
Austrian Academy of Sciences (2021). How fair is the AMS algorithm? (pdf)
Vitt C. (2020). Socio-Technical Analysis of the so-called AMS-Algorithm
Dell'Elce, A., López-Navarro Hochschild, J., Smajević, A., Panezi, A., Garcia Mexia, P. Living with the Algorithm - Toward a New Social Contract in the Age of AI
Allhutter D., Cech F., Fischer F., Grill G. (2020). DER AMS-ALGORITHMUS Eine Soziotechnische Analyse des Arbeitsmarktchancen-Assistenz-Systems (AMAS) (pdf)
Allhutter D., Cech F., Fischer F., Grill G., Mager A. (2020). Algorithmic Profiling of Job Seekers in Austria: How Austerity Politics Are Made Effective
Mager A., Allhutter D. (2020). An algorithm for the unemployed?
Lopez P. (2019). Reinforcing Intersectional Inequality via the AMS Algorithm in Austria
Austrian Academy of Sciences (2019). AMS algorithm on trial (pdf)
Cech F. (2019). Accountability, Bias and Transparency (pdf)
https://algorithmwatch.org/en/austrias-employment-agency-ams-rolls-out-discriminatory-algorithm/
https://dig.watch/updates/discriminatory-employment-algorithm-towards-women-and-disabled
https://www.ippi.org.il/algorithmic-bias-in-the-public-sector-a-view-from-austria/
https://netzpolitik.org/2019/streit-um-den-ams-algorithmus-geht-in-die-naechste-runde/
https://netzpolitik.org/2020/oesterreich-ams-datenschutzbehoerde-stoppt-jobcenter-algorithmus/
Page info
Type: System
Published: February 2023
Last updated: August 2024