From Rotterdam to South Africa: Reflections on Presenting at SACAIR 2025

Starting by giving back

By Marlon Kruizinga

The paper we presented centres on something we have been grappling with throughout our time in Lombardijen. Neighbourhoods like this one have often been the subject of extractive research. Academics arrive, collect data on problems they assumed were there, and leave, having done little to address what residents actually care about.

We believe researchers in these contexts need to start by giving back. That means working alongside residents on existing community initiatives before ever asking them to contribute to a research agenda. It takes longer. It is messier. But it builds something that more efficient approaches cannot: genuine trust, and a research process that is actually accountable to the people it involves.

The value of this approach became clear in hindsight. We had entered Lombardijen expecting to focus on the impact of surveillance technologies. What we found, through sustained conversations with residents, was that their most pressing concerns were about something else entirely: litter and the neglect of public spaces. We also had not anticipated the pride residents feel in Lombardijen's greenery, something that could easily have been missed and would have left our portrait of the neighbourhood incomplete, or even unfair.

Being at SACAIR showed us how global these questions are. Researchers at the conference were working on tools to amplify rather than sideline underrepresented voices, from language models for so-called low-resource languages like Zulu and Afrikaans, to AI literacy programmes embedded within local organisations in low-income communities. It was energising to see so many people thinking carefully about who AI is actually built for.

Lost in translation, and what we found there 

By Nanou van Iersel

We did not quite know what we were walking into. We arrived to present work rooted in philosophy, law, and the social sciences. The conference turned out to lean heavily towards computer science, far more than we had expected. Many presentations took us well outside our comfort zone, into territory marked by equations, system architectures, and algorithmic models we were not equipped to follow.

But that initial disorientation became one of the most valuable parts of the trip. Being surrounded by a research community so different from our own gave us a direct sense of what computer scientists are currently working on, what drives their thinking, and what excites them. We befriended some wonderfully patient researchers who helped translate the technical material for us. In return, we found ourselves explaining concepts like epistemic injustice, which feels almost second nature in our field, but which prompted genuine puzzlement from people who work in quite different ways.

One moment captured the cross-disciplinary tensions perfectly. During a legal presentation on deepfake pornography, the presenter closed with a joke: if it were up to them, they would burn this AI to the ground, and regulation is merely second best. Several of us laughed. Some computer scientists in the room did not. For them, it felt like an attack on their entire field rather than a critique of a specific harmful application. It was a small moment, but a revealing one: a vivid illustration of how differently our disciplines read the same remark.

SACAIR taught us more than we expected, not just about AI, but about the cultural textures of academic fields, about how easily researchers from different traditions speak past each other, and about how much there is to gain when we make the effort to translate.

Between academic blocks, and beyond 

By Majsa Storbeck

Presenting at SACAIR pushed us outside our familiar academic and geopolitical frames, which turned out to be exactly what our research needed.

Coming from a largely EU-based context, we suddenly found ourselves in conversations shaped by AU (African Union) and BRICS (Global South emerging economies) perspectives. As such, discussions not only cut across disciplines but also across what we might call “academic blocks”: different assumptions about what technology is today, how to study it, and what values should guide it. It was refreshing, stimulating, and often confronting, in the best possible way. The questions we were asked were new, stimulating and thought-provoking.

Outside the conference rooms, I spent time exploring the country through walking tours and museums. I walked through Cape Town with a local guide, visited the Apartheid Museum, and joined tours through Johannesburg and Soweto. What struck me most was the directness with which people spoke about their history, without filters, without euphemisation. There seems to be a shared conviction that everything must be discussed, collectively and publicly, precisely so that the violent and discriminatory structures of Apartheid cannot quietly reassert themselves.

The Constitutional Court Museum left a particular impression. Standing in a former prison where both Mandela and Gandhi had been held, and then stepping into the Constitutional Court opened by Mandela, filled with art celebrating dignity and equality, was genuinely moving. Our guide described it as the country's journey from tragedy to triumph, and it felt exactly like that.

Through all of this, I found myself thinking about Ubuntu, the philosophy captured in the phrase "I am because we are", and about how this has shaped South African approaches to Restorative Justice through the work of Desmond Tutu and Nelson Mandela. The idea that a society can only move forward by confronting harm openly, and that healing matters more than punishment, will strongly shape the next phase of my research as we turn toward questions of green crime and harm, approached explicitly from a restorative justice perspective.

In that sense, the trip marked something real: the closing of one chapter and the opening of another.

Marlon Kruizinga, Majsa Storbeck, and Nanou van Iersel are PhD researchers with the AI MAPS project at Erasmus University Rotterdam. Their paper, co-authored with Michaël Grauwde, was presented at SACAIR on 3 December 2025.