The AI race we whould be slow walking: from reflection to introspection.

One of the objectives of my sabbatical in Africa is to revisit my past thoughts, public contributions and academic (non-)achievements, and weave them together into new perspectives and pathways for the future. This piece is a first attempt to do just that, against the backdrop of growing concerns about the global AI race and in preparation for my contributions to the Fak’ugesi Festival in Johannesburg.

The world is again preoccupied with AI. This time because of the well-orchestrated framing by Big Tech that AI has become so advanced that it will destroy humanity. A few years ago, their framing was more nuanced: that the advancement of AI was so powerful that we needed to slow it down, or even pause it. And before that, we saw other, more subtle framings: we are too big to fail, so just support or join us [1]. Those who have worked in AI for long - like myself - will recognize that Big Tech was already, shortly after the 2008 financial crisis, shaping the AI narrative around market logic: competition, speed, efficiency, profit and power.

Big Tech’s current frames have become more explicit and extreme. Most will see this as a natural response to AI becoming more powerful and reaching superhuman levels of intelligence. But I - like many others - believe it has more to do with Big Tech realizing that its objectives are increasingly slipping out of reach. While their repeated fear appeals may affect some people for some time, these appeals are also likely to lead to fatigue and defensiveness, making it harder to turn people into passive AI consumers. At the same time, Big Tech is facing growing geopolitical competition and increasing calls for regulation, accountability and democratic legitimacy.

So, while the concentration of AI power and its misuse by Big Tech and other actors remain worrying, given people’s susceptibility to fearmongering and brain-hijacking [2], there is also reason for hope. It seems increasingly difficult for Big Tech to sustain and sell the idea that it is building modern Towers of Babel, i.e. all-knowing AIs that will serve humanity. And it seems increasingly possible that confusion, chaos and competition may eventually bring these modern Towers of Babel down, paving the way for more diverse and meaningful forms of AI, just as the biblical confusion of languages gave rise to new languages and cultures [3].

The downfall of these modern Towers of Babel may not be painless. It may result in global crises for which Big Tech may already be pre-emptively avoiding responsibility and accountability by calling for the very rules, regulations and policies they for long have avoided, undermined or bent [4]. These crises will be the price we have to pay for collectively being ignorant, oblivious to, and sometimes even supportive of, the coalescence of traditional and new forms of political authoritarianism, economic enslavement and social entrapment. But these crises will not end humanity, as humans and societies are resilient [5]. At most, they will shake societies.

But, rather than becoming preoccupied with Big Tech’s AI frames and the crises they foretell, perhaps we should finally start thinking more seriously about what comes next. However difficult it may be to imagine, at some point current Big Tech powers and practices will give way to new times and new generations [6]. Are we - and in particular our governments - prepared and capable of filling those spaces when the opportunity arises? Have they learned - or are they learning-from the past, beyond repeating the importance and need for AI regulation and accountability? Will they be able to build AI in fundamentally new ways that genuinely benefit people and the planet?

I am afraid not, if I speak from 35 years of academic experience in AI, 15 of them devoted to AI for social good, and 5 years of working intensively with governments, civil society and citizens. My experience with the Civic AI Lab is particularly telling. I founded the lab as a public-public AI research lab, with the ambition to develop frameworks, algorithms and civic awareness that promote equality rather than reproduce or deepen socioeconomic disparities [7]. I did so under the assumption that national and local governments, together with the civic sector, would be more inclined than the private sector - and in particular Big Tech - to fundamentally rethink and rebuild AI, for the public good and planetary wellbeing.

Looking back, however, the Civic AI Lab was increasingly expected to accommodate the very logic it sought to challenge: competition, speed, profit and power. It was this, more than the behavior of Big Tech itself, that surprised me, and forced me to reconsider my assumptions about governments. Nevertheless, the lab remained faithful to its mission, centering compassion, care, love, and individual and social wellbeing.  Confronting governmental (and occasionally academic) institutions with their own logic - whether they asked for it or not - gradually became the lab’s core business. Even when over time, it clearly came to be seen less as a trusted partner or model to follow, and more as a nuisance.

These experiences were reinforced by my work with the Dutch State Commission against Discrimination and Racism, established in 2022 in the aftermath of the childcare benefits scandal and Black Lives Matter. The Commission investigated discrimination and racism in Dutch society, including how the government itself uses data, algorithms and profiling [8]. What I encountered there echoed my experience with the Civic AI Lab: despite growing awareness of the risks, governments nationally and locally continued to adopt many of the same market-driven logics, focusing on efficiency, costs, and control rather than on empowering citizens, communities, civiel society and the public good.

Big Tech behaving like Big Tech is one thing; governments reproducing the same logic is the deeper problem. At the most fundamental level, and across the board, I experienced a widespread absence of institutional reflection and an almost complete lack of institutional introspection. I have seen some institutions engage in reflection, and some even in self-reflection, examining their own role and influence in past AI related decisions, actions and practices. But introspection, which requires institutions to critically examine their inner thoughts, beliefs, assumptions and practices, was practically absent. Instead I experienced an abyss of institutional hubris between reflection and introspection [9] .

As I have argued before, reflection is a precondition for introspection, introspection for imagination, and imagination for meaningful transformation of society - politically, economically, socially or technologically [10]. We cannot design, develop and deploy AI that truly benefits people and the planet without going through these stages. This is why, more than a decade after I first started speaking publicly about and encouraging reflection [11], I find the continued absence of reflection and introspection - the first and most basic stages - across governments (and beyond) so worrying. When modern Towers of Babel collapse, we risk filling the resulting void with principles, practices and products that offer little that is fundamentally new.

This leaves me wondering: is the widespread absence of institutional reflection and introspection simply a feature of human nature, and therefore universal, or is it shaped by the particular societies we have built in the West? Once institutions discover powerful tools that promise efficiency, effectiveness, competition and growth, do they inevitably become drawn towards them? Or has market thinking become a vortex - a centripetal force - that has permeated the private, public and civic sectors so deeply that it shapes not only what we do, but also our ability to reflect, introspect, imagine and transform? If the latter is true, perhaps the search for new forms of AI must begin with finding spaces where individuals, communities and institutions can step outside that vortex and imagine differently.

My current sabbatical in Africa is a deliberate first exploration of such spaces - places, communities, initiatives and institutes where reflection, introspection, imagination and transformation are lived as a way of ensuring balanced social and ecological co-existence rather than simply talked about. I have planned to engage with individuals and communities who make time to revisit their histories, question their assumptions and examine their practices, and with institutions that have the courage to do the same. And I hope to listen, observe, learn and understand what makes these spaces possible, what sustains them, and what they might teach us about collaborating and developing different ways of relating to AI, society and the natural world [12].

This is why I am excited and looking forward to participate in events such as the Fak’ugesi African Digital & Innovation Festival in Johannesburg. The festival is structured around three themes: Memory Engines, exploring heritage as a living system; Imagination Engines, interrogating AI, creativity and agency; and Future Engines, tackling cities, climate, systems and the future of work. In many ways, this reflects the progression from reflection to transformation, creating a space to explore whether AI can be engaged with in fundamentally different ways, grounded in the "ethics of life" instead of "ethics of the market" [13].

Africa is vast, diverse and deeply heterogeneous. I do not see it as a homogeneous alternative to the Western, market-driven world, nor do I want to romanticize it as somehow untouched by market logic or AI technology. It is deeply entangled with the same global economic and technological forces, while also carrying its own histories and ongoing challenges of colonialism, oppression and dictatorship - one of the reasons I currently cannot visit the region where I was born. Yet precisely because of these challenges, and because of its large, young and creative population, Africa offers an important space to explore different ways of relating to community, society, ecology and technology.

Selected public contributions (mostly in Dutch)

[1] The Struggle Against Racist Machines. Interview, Filosofie Magazine, 2018.

[2] Our Brains Have Been Hacked. Here’s What We Can Do About It. Essay, Vrij Nederland, 2020.

[3] How Do You Know What You Don’t Know? Lecture, Paradiso, Amsterdam, 2015.

[4] How We Can Make Artificial Intelligence Make the World Better for Everyone. Essay, Vrij Nederland, 2020.

[5] Let’s Have FAITH: Fair AI Technology for Humanity. Public Lecture, Waanzin Festival, Utrecht, 2018.

[6] Every citizen should be digitally literate and numerate. Chapter, Towards a Cabinet of Social Justice, 2021.

[7] Civic AI Lab launch with Ruha Benjamin. Public Launch, Amsterdam, 2020

[8] Principles for Profiling. Report, State Commission against Discrimination and Racism, 2026.

[9] Staying Real, Staying with the Trouble. End-of-Year Lecture, Kenniscentrum Ongelijkheid, Amsterdam, 2025.

[10] Infinity, Kaleidoscopy and Society: On Reflections of AI. Inaugural Lecture, Amsterdam, 2023.

[11] What We Can Learn from Computers about Discrimination. Mind Shakes IMHO, Amsterdam, 2015.

[12] Entangled Futures. Editorial, AI & Society, 2026.

[13] Paolo Freire, Pedagogy of the Oppressed, New York: Herder and Herder, 1970.