Growing up in Minna, North-Central Nigeria, Samuel Segun dreamed of building sci-fi robots. But after studying philosophy, his focus shifted from Terminator-style apocalypses to the immediate dangers of everyday algorithms. From developing Afro-ethical AI frameworks to crafting algorithmic auditing guidelines for INTERPOL and UNICRI, Segun tells JOHN ADOYI in this episode of My Life in Tech that the most urgent AI safety challenge is not a futuristic machine uprising, but fixing the systems running right now.
Much of the world’s conversation about artificial intelligence (AI) safety begins with a question: what happens if the technology becomes too powerful for humans to control?
Samuel Segun, an AI/ML product leader and AI safety researcher, was once interested in that question too.
“When I wanted to do stuff around AI safety, I was more focused on the very existential part,” he says. “Things like AI systems taking over the world, things you find in Terminator and iRobot.”
Then his research took him somewhere else.
He began to see that some of AI’s most consequential problems were already here: algorithms that discriminate against minority communities, facial recognition systems that perform poorly on people of colour, and automated decisions whose consequences can be difficult to challenge.
For Segun, the question became less about whether machines might one day take over the world and more about who gets protected, who gets overlooked, and whose values are reflected when those machines are built.
But Segun’s interest in technology started long before he had a name for any of this.
Growing up in Minna, Niger State, in north-central Nigeria, Segun was a curious child who loved reading and building things. His father kept a library at home that Segun would explore, and he brought home Video Home System (VHS) tapes of Discovery Channel and National Geographic programmes. Later, when the family installed a satellite dish outside their home, Segun became fascinated by programmes such as How It’s Made and MythBusters.
“I owe a lot of what I am to the forward-thinking nature of my dad,” he says. “He would always encourage us to read and take an interest in things.”
At Government Secondary School, Minna, Segun found someone who shared his fascination with the future: Nasiru, his friend.
“We would look at BMW [cars] or some of these [other] futuristic [vehicles] in magazines,” he says. “Then we would just start talking about designs and things we could build. He was very good with the mechanical stuff, so he would do the robotics part, and I would come up with lots of ideas of things we could build.”
But Segun’s curiosity eventually moved beyond building things. He became interested in questions about the world, the mind, and human existence.
That led him to study philosophy at the University of Calabar, a public tertiary institution in Cross River, southern Nigeria, in 2008.
“I chose philosophy because I felt it could answer some of the very curious questions I would ask,” he says. “Some questions I would ask, I’d just be told, ‘Hey man, don’t ask those questions, you’re going to get yourself confused or mad.’ But I kept asking.”
His questions soon took him beyond philosophy’s traditional boundaries. He spent time with medical students, discussing diseases and imagining medical devices they could build. His undergraduate research focused on neurophilosophy and the problem of identity, leading him to questions about consciousness and artificial life.
By the time he began his master’s degree, those questions had become the focus of his research.
From questions to products
When Segun graduated from the University of Calabar in 2012, he founded Uptrend Media, a startup that helped early-stage tech companies research their markets, develop minimum viable products (MVPs), and work out how to take their ideas to market.
The idea grew from something he had been doing at university. After watching how polls were used during American elections, Segun began conducting his own surveys during student union elections. He distributed questionnaires across campus and worked with a friend in the mathematics department to analyse the results and publish predictions.
“We could predict with a certain degree of certainty who was likely to win the SUG [student union government] elections based on the feedback we got from our survey,” he says.
After he graduated, he began thinking about how that experience could become a service for startups.
While running Uptrend, one of his clients introduced him to his next job.
Academix.ng, a Nigerian platform that digitises academic research and resources, was trying to solve a problem Segun knew well: accessing Nigerian academic research. Its founder was digitising dissertations and journal articles while still figuring out what the product should become.
Segun was helping the company with its campaigns and onboarding process when he suggested that its understanding of what researchers needed was incomplete. He knew the research process from his own academic work and had already begun publishing.
The conversation led to an offer to join Academix as head of its product team in 2016.
“I handed over Uptrends Media to my co-founder and joined Academix,” he says.
While there, Segun got his first real exposure to AI. Academix hosted the then-Facebook Developer Circle, whose team would set up in one of the meeting halls and play videos from Facebook Artificial Intelligence Research, Facebook’s AI research team.
Some of it was far beyond what Segun could understand at the time. But it caught his attention.
“I was like, I’m interested in this AI stuff, and I want to do it, but I don’t know how to do it.”
In 2017, he left Academix and joined Univelcity, a technology training institute, where he worked in business development and product management while supporting its fellowship programme.
But AI gave him something new to think about. He said he began researching the technology online and thinking about its ethical implications. His background in philosophy had already made him interested in ethics, and he began considering what those questions might mean for a technology he wanted to understand.
According to him, he began writing PhD proposals to universities, looking for a way to study the field. In January 2018, he left Nigeria for South Africa to begin his PhD at the University of Johannesburg.
Rethinking AI safety
When Segun began his PhD, he focused on existential risks in AI safety.
“My initial approach to it when I wanted to do stuff around AI safety was that I was more focused on the very existential part,” he says.
As he researched further, he found problems already affecting people: algorithmic bias, facial recognition systems that struggled to recognise people of colour, and tools such as Correctional Offender Management Profiling for Alternative Sanctions (COMPAS), which are used to predict whether a person will commit another crime.
“I just realised that there were a lot more pertinent problems that you could actually find,” he says.
His questions became more specific. Could fairness be represented mathematically? Could non-bias and transparency be represented in a way that could actually be applied to an AI system?
At the University of Johannesburg in 2019, in South Africa, Segun worked with researchers and engineers at the Institute of Intelligence Systems while developing his PhD research. That work led him to another question: if people in different societies understood ethical questions differently, should AI systems also account for those differences?
This became the focus of his thesis, “Constructing an Afro-ethical Framework for Autonomous Intelligence Systems.” He explored whether ethical principles rooted in African and other cultural contexts could help shape AI systems for the societies in which they would operate.
The idea also informed a 2020 paper, Critically Engaging the Ethics of AI for a Global Audience, in which he argued that cultural differences in how people understand ethical questions around technology should be considered when building AI systems.
As his work developed, Segun says it began attracting attention within the university.
“I think at that point it became clear to me,” he says. “I was contributing something very essential.”
After completing his PhD in 2020, Segun joined Stellenbosch University, South Africa, as a postdoctoral fellow, where he continued his work on computational and data ethics.
Building guardrails for AI
In January 2022, he joined Praelexis AI, a South African machine learning research company, as an artificial intelligence and machine learning (AI/ML) product manager.
The company worked on projects for multiple clients, including banking analytics, and Segun saw firsthand how AI systems could go wrong, according to him.
“I could see the role of algorithmic auditing and why it was important to audit algorithms and understand where systems could go wrong,” he says.
He also saw how models were tested and stress-tested, and how factors such as location, gender, age, and race could affect their decisions.
At the same time, Segun worked as an AI Innovation and Technology Consultant for the United Nations Interregional Crime and Justice Research Institute (UNICRI), a United Nations research institute focused on crime prevention, criminal justice, and the responsible use of emerging technologies, in collaboration with INTERPOL.
UNICRI and INTERPOL started a project to develop a tool for law enforcement. Segun says he was one of the consultants on the project. He developed guidelines for auditing algorithms, assessed organisations, and worked on readiness and risk assessments.
Part of his job was to translate principles such as fairness, explainability, and robustness into practical steps that law enforcement agencies could use.
“We would always have interactions and meetings with Interpol teams,” he says. “And they would say, ‘If we’re going to be operationalising this, it might be difficult for police chiefs.’”
The experience showed him the gap between discussing AI ethics in academic settings and actually implementing it inside institutions.
“I think more importantly, the work that I did with UNICRI and Interpol, and much later with the United Nations Office of Counterterrorism, taught me how to scale ideas,” he says.
While still working as a consultant for UNICRI, Segun returned to product management. In June 2023, he joined Evolve Credit, a Nigerian financial services infrastructure company, where he worked on its expansion into North America and supported the development of AI-enabled tools with the machine learning team and CTO.
Making AI governance practical
In June 2024, Segun joined the Global Centre on AI Governance as a Senior Technical Programme Manager, where he led research on AI safety, including the evaluation of frontier language models and the categorisation of AI risks. He also led work on Toward an African Agenda on AI Safety, a paper that mapped risks requiring particular attention on the continent.
“It is a unique paper because it mapped out close to 15 risk categories that are unique to the continent,” he says. “It is one of the things I kind of hold dearly.”
The work reinforced his view that Africa faced a significant gap in the capacity and knowledge needed to prepare for AI’s impact.
“You can see this with how the algorithms on social media platforms are reinforcing certain beliefs or stirring kids into a particular way and all that,” he says. “It’s not just any technology; it’s very consequential technology.”
For Segun, the challenge was not simply access to AI technology. Governments and institutions also needed the expertise to understand AI, assess its risks, and make informed decisions about how it should be used.
“We need to support policymakers, practitioners and all that with the right tools, right knowledge of the technology, so that when they’re building policies, it’s policies that are sound, well-informed, [and] take into account the local challenges.”
That meant giving policymakers enough technical understanding to identify risks and demand safeguards for AI systems that could have significant effects on areas such as healthcare and education.
“How do we train policymakers to really understand the implications? How do we train them to understand technology enough to say, ‘Hey, these are some of the hard lines. We would require auditing of a model’?” he says.
He saw this as a question of institutional capacity, much as it had been with data protection.
“The same way we have institutional capacity for data protection, we’ll need to have that more for AI,” he says.
That concern with making technical knowledge useful beyond specialist circles also shaped another project. While at the Global Centre, Segun founded Algorithmic Review, a publication focused on emerging technologies and voices from the Global Majority.
“The idea was to ensure that another person could pick up an article, read it without all the jargon, and still understand the implications of that technology on them and what they need to do,” he says.
He wanted people to understand not only what emerging technologies could do, but how they might affect their lives.
Building for the long run
Now, Segun is back to building frontier technology himself.
In 2025, he co-founded Quelox Labs, an industrial AI company based in Vancouver, Canada. Its first product, Factori, is designed to connect to existing factory equipment and give manufacturers real-time visibility into their operations.
The company is using that data to develop tools for automating processes, diagnosing problems, and predicting when machinery needs maintenance before a breakdown becomes costly.
For now, Segun says Quelox is focused on launching across multiple factory floors and raising its pre-seed round as it grows.
But for him, building Quelox is also part of a longer journey of learning by doing. He believes more people should experiment with ideas, even when they are unsure where those ideas will lead.
“Sometimes, just build things for the sake of building them,” he says. “They don’t have to be grand or become a funded startup. And it’s okay to build something and fail. There’s so much to learn from it.”
True scale demands moving beyond surface-level integrations to robust execution. We’ve filtered the noise out of Moonshot 2026, optimising the conference strictly for high-calibre connections between startup founders, global financial operators, enterprise leaders and individuals rewiring Africa’s technical frameworks. Get 20% off Early Bird tickets for a limited time.
