Logo:
Bilateral-AI/News & Events

BilAI Workshop: Towards well-behaved AI Systems at the AI Festival at TU Wien

A workshop situation with researcher Marta Sabou standing in front and speaking and a group of participants listening and watching her.

15.12.2025

During this year’s AI Festival at TU Wien in early December, the Bilateral AI cluster of excellence convened a well-attended and intellectually engaging workshop titled “Towards Well-Behaved AI Systems: In Search of Alignment.”

The BILAI session, led by Kees van Berkel from TU Wien and Marta Sabou from Vienna University of Economics and Business, both key researchers in BILAI’s “ethical AI systems” research module, brought together around 90 participants from diverse backgrounds, including industry partners, start-ups, legal experts, university researchers, and students. 
Motivated by initiatives such as UNESCO’s AI ethics recommendations and Austria’s Digital Humanism framework, the workshop highlighted the need for interdisciplinary collaboration among developers, computer scientists, policymakers, social scientists, ethicists, and the public.

Interdisciplinary Dialogue at the Core

After introductory talks by van Berkel and Sabou on AI alignment, fairness, and auditing methods, participants split into small groups to tackle seven key questions:

  • What is well-behaved AI?

  • How can it be achieved?

  • What technical challenges exist?

  • Who is responsible?

  • What policies are needed?

  • How can society be engaged?

  • What role should universities play?

Reflecting on these discussions, Isabella Cissell, BilAI PhD student and discussion host, highlighted the collaborative dynamics of the workshop:

"It was interesting to see how quickly a group of people from various backgrounds were able to find common ground on the topic of well-behaved AI. My group was tasked with defining this concept, and we were able to identify and discuss several interpretations of what well-behaved AI might mean, especially in comparison to other concepts we often hear discussed, like alignment, transparency, and trustworthiness.”

Key Outcomes

Discussions revealed the complexity of creating well-behaved AI and emphasized:

  • Shared responsibility across stakeholders

  • The importance of EU AI Act compliance and policy clarity

  • Auditing, transparency, fairness, and explainability as core challenges

  • Universities as facilitators of interdisciplinary collaboration

The workshop underscored the need for ongoing, multi-stakeholder engagement to develop AI systems that are both powerful and aligned with societal values and expectations.

 

Back to news
Subscribe for news