Improving response to risks of discrimination, bias and intolerance in automated decision-making systems to promote equality.

ongoing

Public authorities’ uptake of AI and other algorithmic tools in data collection, processing and services is a rapid and, most probably, a non-reversible new reality. Globally, many governments use automated decision support tools or even automated (algorithm-based) decision-making (ADM) systems in their everyday operations. ADM is the process of making a decision or creating other legally relevant output through automated means with limited or no human involvement. The ADM systems may process large amounts of information, identify patterns, make predictions, or take various types of actions based on predefined criteria. The general objective of the project is to enhance the capacity of public authorities and their co-operation partners in Estonia and Lithuania in addressing risks of discrimination and bias in ADM systems. This will be achieved by in-depth research, developing support materials, trainings and a media campaign. This project seeks to contribute to prevention of discrimination and gender inequality within the European Union context by ensuring that the ADM systems used in the public sector are fair and non-discriminatory. The key activities include: 1. Research on gaps in policies and legal frameworks on AI in Estonia and Lithuania together with mapping of potential risks of discrimination and bias in ADM systems deployed by Estonian public authorities; 2. Development of guidelines, impact assessment checklist, training materials and online course for public sector employees and system developers to address potential risks for discrimination and bias in design and deployment of ADM systems; 3. Capacity building among project stakeholders and target groups through trainings, study trips, international conference and an academic workshop; 4.Raising public awareness on the risks of discrimination and bias in public sector ADM systems.