February 6 09:30-12.30

LUX Auditorium (Lower)

09.30 Introduction by Ericka Johnson, WASP-HS Graduate School Director

09.35 Johannes Widegren:

Archives + Sápmi + AI = ?

Abstract: Why should Indigenous peoples have special rights? When Sweden needs more critical minerals, timber, or hydropower, why should we let the Sámi stand in the way? The decision of whose needs to prioritize depends on which narrative we tell about the past. Swedish society has regarded the Sámi as an exotic people group, very useful for tourism but with little relevance to a modern Swedish society. Meanwhile, the narrative told by the Sami themselves, speaks of an ingenious people which has managed to survive under enormous pressure from state-backed exploitation, paternalism and colonialism. As our collective memory increasingly relies on what can be accessed online, it is of utmost importance that the Sami narrative is discoverable, especially when biased AI tools are used to make archives searchable – or the colonial narrative may again be reinforced. That’s why I collaborate with the Sámi periodical Samefolket to explore if we can use AI tools to their advantage to make their entire digitized news archive discoverable for Sámi and non-Sámi users alike.

10.05 Nicolas Audinet de Pieuchon:

Countering Bias in AI Methods for Text Data

In this talk I will give an overview of my research journey so far in the hope that some of you will find it useful! The purpose of my research is to develop ways to counter bias in AI methods for text data. More concretely, I have been focusing on creating ways to mitigate, ignore and remove unwanted information in text datasets. In my current project I’m using tools from the mechanistic interpretability community to try to “blind” a language model so that it ignores certain variables (e.g. gender information when predicting profession from biographies). In previous projects I’ve tested methods to mitigate biases from LLM-generated annotations with a small quantity of human annotations, concluding that these methods are sound and should be used, and attempted to use LLMs to remove unwanted information from text by rewriting it, concluding that they are not fully able to do so.

10.35 Coffee break

11.00 Ziming Wang:

A Meat-Summer Night’s Dream: A Tangible Design Fiction Exploration of Eating Biohybrid Flying Robots

What if future dining involved eating robots? We explore this question through a playful and poetic experiential dinner theater: a tangible design fiction staged as a 2052 Paris restaurant where diners consume a biohybrid flying robot in place of the banned delicacy of ortolan bunting. Moving beyond textual or visual speculation, our “dinner-in-the-drama” combined performance, ritual, and multisensory immersion to provoke reflection on sustainability, ethics, and cultural identity. Six participants from creative industries engaged as diners and role-players, responding with curiosity, discomfort, and philosophical debate. They imagined biohybrids as both plausible and unsettling—raising questions of sentience, symbolism, and technology adoption that exceed conventional sustainability framings of synthetic meat. Our contributions to HCI are threefold: (i) a speculative artifact that stages robots as food, (ii) empirical insights into how publics negotiate cultural and ethical boundaries in post-natural eating, and (iii) a methodological advance in embodied, multisensory design fiction.

11:30 Laetitia Tanqueray:

Reframing Robots for Care: Situating Informal Caregivers in the Development of Social Robots

With the looming change in demographics, States are left to find sustainable solutions. With the rise in Artificial Intelligence, it has become a possibility that AI-driven solutions will become a solution, including robots. However, are robots the saviors of care? This talk aims to nuance this, by demonstrating that care is already relying on non-State actors to ensure that people in need of care get adequate support, namely informal caregivers. Therefore, whilst robots may provide care in the future, the reliance on informal caregivers will continue (and possibly increase).

12.00 Sarah de Heer:

When Algorithms Shape Care: AI-Driven Medical Device, Precision Medicine, and Quality under the Right to Health

The research examines whether the conformity assessment procedure under the Medical Devices Regulation and the Artificial Intelligence Act can adequately ensure the quality of AI-driven medical devices in precision medicine. It further investigates the extent to which Notified Bodies are able to assess these devices in a manner that safeguards the quality pillar of the right to health for all segments of society, given the opacity and data-dependency inherent in AI systems.

AI is increasingly embedded in medical devices used in precision medicine, promising faster diagnoses, personalised treatments, and improved clinical outcomes. Yet, alongside these benefits arise significant challenges as regards the fundamental right to health, and particularly its pillar ‘quality’. In the European Union, quality requires health care to be safe and effective. Accordingly, AI-driven medical devices must successfully complete the conformity assessment procedure under the Medical Devices Regulation and the Artificial Intelligence Act, which are overseen by Notified Bodies.

When AI-driven medical devices in precision medicine are trained on incomplete or unrepresentative datasets, they may generate inaccurate diagnoses or inappropriate treatment recommendations, potentially causing serious harm. These risks are not hypothetical, as biased datasets may improve outcomes for individuals who resemble the training data while disadvantaging those who do not match the training data, thereby exacerbating health inequalities. My doctoral research examines these risks through the lens of the right to health, focusing on its quality pillar. Further, my project questions whether the Notified Bodies can meaningfully assess the quality of AI-driven medical devices in precision medicine given their opacity and data dependency.

.