Skip to content

Algorithm used in Austrian employment follows GDPR Article 22, according to court decree

Austrian court annuls data protection agency's prohibition on job-matching algorithm, deeming human oversight adequate to prevent automated decision-making abuses.

Austrian court finds employment algorithm in compliance with GDPR's Article 22 regarding automated...
Austrian court finds employment algorithm in compliance with GDPR's Article 22 regarding automated decision-making and profiling.

Algorithm used in Austrian employment follows GDPR Article 22, according to court decree

The Austrian Public Employment Service's Labor Market Opportunities Assistance System (AMAS) has resumed operations following a court ruling that overturned a previous ban. The system, which categorises job seekers based on their potential employment opportunities, was suspended in January 2021.

The AMAS algorithm calculates integration probabilities for unemployed individuals using personal characteristics such as age, gender, education level, health status, and regional labor market conditions. It assigns job seekers to three categories: high labor market opportunities, medium opportunities, and low opportunities. Each classification determines access to different support measures and funding pools.

Employment advisors play a substantial role in the AMAS system. They independently assess job market prospects and are authorised to correct AMAS values when their evaluations differ. Advisors receive computer-generated "segment supplementary information" providing detailed explanations for algorithmic classifications. They are also required to document disagreements with job seekers regarding market opportunity evaluations and explain their reasoning during mandatory consultations.

The system includes technical safeguards preventing automated overrides of human corrections. Each office employs ombudsmen to handle client complaints about assessment decisions. Additionally, an "AMAS representative" is responsible for conducting plausibility checks when advisors question the system's accuracy.

The court found that the Public Employment Service's training programs instruct advisors to recognise factors not captured by the algorithm. Classification into high-opportunity groups emphasises rapid job placement, while medium-opportunity clients receive intensive support across all available services. Low-opportunity individuals access 90-95% of support measures but are excluded from expensive specialized training programs.

The Austrian Federal Administrative Court's ruling on September 1, 2025, stated that the AMAS does not violate GDPR Article 22 prohibitions against automated individual decision-making. The court extensively referenced the December 2023 European Court of Justice Case C-634/21 involving SCHUFA credit scoring. The ECJ established that automated processing constitutes prohibited decision-making under Article 22 when it plays decisive roles in subsequent choices, regardless of nominal human involvement. However, the Austrian court distinguished AMAS from the SCHUFA case, citing that employment advisors possess real decision-making authority rather than perfunctory oversight.

On July 8, 2025, European authorities endorsed GDPR record-keeping modifications. However, there is no information available regarding which organisation has the authority to review revisions of the Austrian Public Employment Service's AMAS algorithm or how many times such reviews have been conducted.

The court's decision overturns a 2020 Data Protection Authority ban that had effectively shut down the controversial algorithm. Despite concerns about potential unconscious bias in advisors, the court's decision dismissed these concerns.

The resumption of the AMAS system is a significant development in the Austrian job market, providing a tool for employment advisors to better assist job seekers in their search for employment.

Read also:

Latest