On Friday, May 16th, the first alignAI Seasonal School concluded with a dynamic round of presentations by the Doctoral Candidates (DCs), marking the culmination of an intensive week of getting to know one another, learning and collaboration. Hosted at the Garching office of the Institute for Ethics in Artificial Intelligence, this session brought together the alignAI DCs to pitch their emerging ideas on aligning LLMs with human values and societal wellbeing.
Three interdisciplinary working groups took the stage, each centred around a distinct, real-world use case. From mental health to education and online news consumption, the DCs demonstrated how values, legal considerations and cutting-edge technology can come together to build ethical and human-centred AI systems.
Group 1 tackled the sensitive challenge of mental health support for parents of children with suspected psychological conditions. Their solution? An Agentic-AI companion, a legally and ethically informed chatbot to guide families through complex care systems. With a strong focus on coordination with their peers across the alignAI network and beyond, and transparent legal frameworks, this group exemplified how AI could empower rather than replace human actors.
Group 2 focused on AI in education, proposing a game-based learning environment to foster both digital literacy and ethical reflection among school-aged children. Their project emphasised iteration and co-design, drawing on values derived from participatory research and tying technical development closely to societal goals.
Group 3 explored the increasingly relevant terrain of online news consumption, analysing how AI can assist both journalists and readers while mitigating risks like misinformation and loss of transparency. Their reflections included legal insights into GDPR, the AI Act and the DSA, alongside a case study of Italy’s Il Foglio AI, a recently published AI-generated newspaper, which raised timely questions and concerns about automation, editorial responsibility and human oversight.
Each group not only outlined their technical and conceptual frameworks but also presented concrete collaboration strategies across the Doctoral Network, highlighting shared design goals, legal-ethical validation cycles and integrated LLM development.
With the final presentations and the conclusion of the first seasonal school, the DCs departed to the next phase of their joint journey: Getting to work, collaborating and designing aligned LLMs.