The information displayed in the AIM should not be reported as representing the official views of the OECD or of its member countries.
The Austrian Public Employment Service’s ChatGPT-powered chatbot, Berufsinfomat, has been criticized for reproducing gender stereotypes by recommending science programs to men and humanities to women. Developed by AMS and AI firm Goodguys to assist job seekers, the tool has sparked user complaints and prompted efforts to correct its discriminatory outputs.[AI generated]
Why's our monitor labelling this an incident or hazard?
The chatbot is an AI system explicitly mentioned as using ChatGPT to generate personalized employment advice. The event involves the use of this AI system, which has directly led to biased and discriminatory recommendations based on gender, a violation of fundamental rights and labor rights. The harm is realized as users receive prejudiced guidance that could affect their career choices and opportunities. The AMS acknowledges the issue and is working to mitigate it, but the incident of bias and harm has already occurred. Therefore, this event fits the definition of an AI Incident due to realized harm from the AI system's outputs.[AI generated]