Meaningful Human Control Center

When systems have autonomous capabilities, meaningful human control is an issue: these systems may act too fast, or act based on data or processes that are too complex to understand and monitor, making it unclear who is morally responsible in case of errors. This issue is exacerbated when these systems are powered by AI (e.g. machine learning algorithms).  More and more decisions previously made by humans are being placed in the hands of artificial systems. The complexity and capabilities of these systems challenge our traditional notions of control and responsibility. This leads to increasing worries about the role and responsibility of relevant humans around these systems. The principle of meaningful human control is all about designing and engineering systems in a way that relevant humans (users, bystanders, designers, organisations, policy makers) have the right amount of knowledge, time, and responsibility to maintain appropriate influence over such systems with autonomous capabilities.

The centre for meaningful human control connects academics and practitioners aiming to conceptualise, design, implement, and assess systems under meaningful human control. We strive to be a lighthouse for collaboration among multiple stakeholders, while leveraging interdisciplinary expertise, existing initiatives at TU Delft, and an international network of collaborators at the forefront of research and practice on meaningful human control. We consolidate TU Delft’s unique expertise on meaningful human control, including the AiTech and the NWO project “Meaningful Human Control over Automated Driving Systems”, and expand towards a broad network of collaborators and practitioners.