Germany: AlgorithmWatch’s study “From risk mitigation to employee action along the Machine Learning Pipeline”

On 16 March, the German NGO AlgorithmWatch, with the support of the Hans Böckler Foundation, published a study advocating that workforce representatives should be more closely involved in the deployment of AI and automated decision-making systems, from the design phase through to their deployment. The authors want to go further than either current legislation, which puts forward a risk mitigation approach, or “trade union discussions focus[ed] on addressing … ethical standards”. “… a step toward more practice- and process-oriented approaches is urgently needed” by “empower[ing] employees in the face of algorithmic decision-making”, via “transparency requirements, support for co-participatory governance approaches, or funding for ADM-focused training programmes for employee representatives”.