AI governance
Abraham Mhaidli | Yixin Zou | Max Planck Institute (Germany)
03/12/2026 12:00
Sala Descubre/Online Teams. Edificio 8E. Acceso J - 4ª Planta | Universitat Politècnica de València. 46022
Title: "AI Ethics: The Limits of Frameworks, and Why We Need to Radically Rethink AI Ethics Enforcement" | Abraham Mhaidli
Abstract
The rapid proliferation of AI has led to concern about the risks and harms these systems bring. In response, there has been a proliferation of AI ethical frameworks and toolkits that aim to help developers produce ethical AI. Despite the existence of these frameworks, many unethical AI systems are still being developed, and it is unclear what effect these frameworks have had on developing ethical AI. In this talk, I critically examine AI ethical frameworks to ask what are their limitations and shortcomings. First, I focus on a recent study where we interviewed developers about whether and how they incorporate AI ethical frameworks into their work. We find that frameworks are often overlooked for legal compliance, and how frameworks and toolkits were often used as evidence of compliance rather than being used to develop more ethical AI. In the second part of the talk, I build on this paper to discuss more broadly the limits of current ethics implementation in CS, and argue that we need to radically rethink how ethics is enforced in software engineering if we truly want to solve the issues that AI, and other tech products, pose.
Bio
Abraham Mhaidli is a Junior Research Group Leader working at the Max Planck Institute for Security and Privacy in Bochum, Germany, where he leads the Technology Harm Prevention Group (aka The Alpaca Lab). His work primarily looks at new and emerging technologies (including artificial intelligence, brain computer interfaces, smart devices, and virtual reality) to understand what are their harms and how to mitigate them. He is also interested in research and technology ethics, meta-science, degrowth, and understanding how capitalism and neoliberalism impact and influence academia. Abraham obtained his PhD in Information from the University of Michigan in 2023. Outside of research, Abraham enjoys watching football, playing board games, cooking, and spoiling his two cats.
_____________________________________
Title: Who Counts as the Human in Human-Centered AI? | Yixin Zou
Abstract
Human-centered artificial intelligence (HCAI) seeks to shift the focus in AI development from technology to people. But the field rests on a question rarely asked: which human? In my talk, I will trace three strands of my group's work, each centering a different one. At the individual level, the human is a lay user: we study which interaction paradigms sustain creativity in human–AI collaboration, and how far AI can go in teaching people to recognize scams or supporting victims through recovery. At the institutional level, the human is a professional inside an organization. Anti-fraud analysts delegate evidence gathering and summarization to AI but keep approve-or-block decisions for themselves, reconciling personal judgment with organizational compliance. At the societal level, the human is a member of the public: we document systematic misalignment between how the public perceives AI risk and how the EU AI Act regulates it, while finding that few people, experts included, calibrate that risk well. Across these cases, different humans want something different from AI. Whose needs and narratives should count in AI design and governance, and how do we navigate these tensions as we move toward agentic AI?
Bio
Dr. Yixin Zou (she/her) is a tenure-track faculty member at the Max Planck Institute for Security and Privacy, where she leads the human-centered security and privacy group. Her research interests span human-computer interaction, privacy, and security, aiming to make technology safer and more equitable for underserved communities. Her research has been recognized with the ACM SIGCHI Outstanding Dissertation Award (2024), the John Karat Usable Privacy and Security Student Research Award (2022), and several best paper and honorable mention awards at top venues such as ACM SIGCHI Conference on Human Factors in Computing (CHI) and the Symposium on Usable Privacy and Security (SOUPS). Her research has also generated broader impacts on public policy, including the rule-making process for the California Consumer Privacy Act. She earned a Ph.D. in Information from the University of Michigan in 2022.
_______________________________________________________________________________
Reunión de Microsoft Teams
Unirse: https://teams.microsoft.com/meet/358298459505446?p=e5fuOItoMK9dAezdLf
Id. de reunión: 358 298 459 505 446
Código de acceso: SY3wG9Py
