As the European Union ushers in a new era of digital governance, researchers and developers face a burgeoning “Accessibility Paradox.” On one hand, the European Accessibility Act (EAA) mandates that digital services and public infrastructures be proactively inclusive, often requiring systems to adapt to the specific needs of users with disabilities. On the other, the EU AI Act introduces stringent prohibitions on “biometric categorisation,” strictly regulating or banning the use of AI to infer sensitive characteristics, such as disability status, for the purpose of data harvesting or model training.
This webinar explores the critical tension between these two landmark regulations. While the EAA pushes for “smart” environments that recognise and respond to human diversity, the AI Act seeks to protect citizens from being “labelled” or profiled by automated systems. This creates a complex landscape for those working in urban planning and digital co-creation: How can we build AI-driven tools that provide seamless accessibility without infringing on biometric privacy? In practice, this paradox is intensified by institutional data deserts, where the lived accessibility needs of vulnerable groups are poorly captured, forcing designers to choose between inference, exclusion, or human mediation.
Drawing on research, this webinar will dissect the legal “gray areas” of AI-mediated outreach. Key discussion points include:
- The distinction between prohibited biometric profiling and legitimate assistive personalisation.
- Strategies for designing privacy-preserving accessibility that prioritise user-led declarations over autonomous system-led detection.
- A clash of rules: Why fixed AI laws and ‘one-time’ consent don’t work for the diverse and changing needs of neurodivergent people
- The challenges of maintaining “human-to-human” connection in digital co-creation tools, such as chatbots, while ensuring they do not become barriers for older or digitally excluded populations.
- Governance tensions and the role of civil society: How can accessible, rights-based AI systems be meaningfully designed and implemented if civil society organisations representing marginalised communities are structurally excluded from regulatory deliberations and co-creation processes?
By reframing these regulatory hurdles as design opportunities, this webinar provides a roadmap for “Double Compliance”, ensuring that the next generation of AI tools is not only technologically advanced but also ethically grounded and universally accessible.
Webinar Outline
Total time: 90 minutes
- 11:00 – 11:05: Welcome Address by Sarah Anne McDonagh (UAB)
- 11:05 – 11:20 When Inclusion Meets Prohibition: The Accessibility Paradox in EU AI Law (Daphne, DBC diadikasia, ALFIE)
- 11:20 – 11:35 Civil Society and the AI Governance Paradox (Alexandros Minotakis, University College Dublin, FORSEE)
- 11:35 – 11:50: The road to Paradox Hell may be paved with the best of intentions (Haris Shekeris, Catalink, ALFIE)
- 11:50 – 12:20 Debate & Open Discussion (Q&A)
- 12:20 – 12:30: Wrap up Sarah McDonagh (UAB)
