
Sign up to save your podcasts
Or


As AI systems become increasingly capable, one question deserves more attention: what happens when they improve faster than our ability to understand them?
In this episode, I explore recursive self-improvement, AI safety, independent external evaluation, and the limitations of guardrails. I share my perspective on why evaluating whether an AI system is safe is not the same as understanding how it makes decisions.
As we move towards more autonomous AI systems, interpretability and observability could become just as important as intelligence itself.
Perhaps the real challenge isn't simply building AI that can improve itself, but ensuring that our ability to understand, audit, and govern these systems keeps pace with their capabilities.
Are we building AI that we can genuinely understand, or are we advancing faster than our ability to remain in control?
In this episode of AI Conversations: Africa in Focus, we tackle a fundamental but often overlooked question: Why are you building that ML system in the first place?
Segun explores the critical distinction between building machine learning models for research versus deploying them to solve real business problems. Too often, teams celebrate marginal gains in accuracy while ignoring whether their models are moving the needle on key business metrics like customer retention, conversions, or cost reduction.
You’ll learn why success in a business setting demands more than a high-performing model—it demands alignment with business objectives, reliable and scalable systems, and a clear experimental approach, such as A/B testing, to measure actual impact.
Whether you're a data scientist, ML engineer, or business leader interested in AI, this episode offers practical insights on how to build ML solutions that matter.
Tune in to discover how to move from model obsession to business impact.
Join this conversation as we unpack the science, politics, and future of AI in this exciting and insightful episode!In this episode ofAI Conversations: Africa in Focus, I dive into one of the most significant developments in the AI landscape — the release of the DeepSeek model series.
This revolutionary model has achieved reasoning capacities comparable to OpenAI's O1 model series but at a fraction of the computational cost. I explore the innovative techniques behind this breakthrough, the implications for the future of AI, and how this could make training language models on consumer GPUs a reality.
In this episode, I also discuss the global reactions to this disruption, including OpenAI's response, Nvidia's market shock, and the European Union’s surprising commitment of €54 million toward building large multilingual language models under the OpenEuroLLM project.
Join this conversation as we unpack the science, politics, and future of AI in this exciting and insightful episode!
Welcome back to the podcast! After a brief hiatus, we’re excited to return with renewed energy, a broader vision, and a new identity. In this episode, we address the reasons behind our break, reflect on our journey so far, and introduce the rebranding of our show from “Applying AI: The AI in Africa Podcast” to “AI Conversations: Africa in Focus.”
We’ll continue to explore how artificial intelligence and machine learning can address Africa’s unique challenges while expanding our scope to include global AI trends, groundbreaking research, ethical concerns, and transformative applications across industries.
Join us as we discuss what’s next for the podcast, how we’re reaching a larger audience, and the exciting changes to come.
Tune in and be part of this new chapter as we bring insightful conversations with experts and thought leaders shaping the dynamic world of AI.
In this episode, we introduce the pressing issues plaguing our nation and what transformative solutions AI could bring to them, from the farmer-herder crisis to unemployment and insurgency. These challenges, deeply rooted in various sectors including agriculture, education, security, finance, and industry, have evolved over time, often unnoticed until they become full-blown crises.
However, the advent of artificial intelligence offers a beacon of hope. AI models can predict future outcomes by harnessing vast datasets containing invaluable patterns and behaviors and offer
AI’s transformative power extends beyond security. It promises to revolutionize businesses, education, and other critical sectors, offering personalized solutions tailored to individual needs and predictive insights that enable proactive problem-solving.
Dr. Bosun Tijani's visionary pursuit of AI technology underscores the urgency for Nigeria to embrace this transformative tool. With 45 pioneers actively engaged in using AI to address challenges across various sectors, the stage is set for a paradigm shift in Nigeria's technological landscape.
And in the end, Nigeria would be actively part of the ongoing AI conversations
in our next episode, I will talk more about how my research projects on the Personalized Intelligence Tutoring System in Education would improve Nigeria’s education outcomes however I was hoping you could think about what it would mean for students to have a system that can provide individualized guidance and support, and assess their understanding of a topic, for example, identifying with deep accuracy areas of difficulty and strengths and offer targeted explanations, feedback, and practice exercise.
Our subsequent episodes would be to invite some grantees of the National Artificial Intelligence Scheme where we’ll go into their specific AI research projects and their potential to tackle Nigeria's most pressing challenges. Don't miss out on the opportunity to be
In this latest episode of Applying AI: The AI in Africa Podcast, Segun welcomes listeners to the first episode of the year 2024, continuing the discussion from the previous episode. The focus is on the challenges of scaling African businesses with artificial intelligence and the importance of identifying the impact before attempting to scale. In this episode, Segun delves into the essential question of whether one can scale the impact of solutions that haven't been implemented yet. He emphasizes the necessity of understanding the "why" behind building an artificial intelligence model for a specific business and explores the key questions that business owners or machine learning engineers should ask themselves or their clients. The overarching theme is that building an AI model is a means to an end, and the episode highlights the importance of defining the goals and purposes the model aims to achieve.
I recently participated in discussions on artificial intelligence in an event I attended in Lagos; Moonshot by techcabal. The discussions on AI centered around scaling the impact of artificial intelligence in Africa and it then occurred to me that what is there to scale when we're just getting started with artificial intelligence and when many organizations and companies are still in the infant stage of building an AI model specifically targeted at a problem. So in this episode, I discuss what exactly must be put in place before we begin to talk about scaling AI impact because scaling AI is just more than the plug-and-paly AI models; it starts from identifying its need in a business, the specific problem it intended to address to sourcing the appropriate data, important features that could make it easy for the model to learn what’s important to its training, evaluations, tunning, deployment measurements and up to enterprise level is a challenge.
Businesses now revolve around the use of technology and as far as we're concerned artificial intelligence is at the center of this business revolution so in this newest episode of Applying AI: The AI in Africa podcast, the host discusses how business owners and managers can leverage on AI to achieve the most results in their business. The host also discusses the myths about artificial intelligence in Business and why they are unfounded; Also considers exactly which area of the business should business owner deploy AI systems and three fundamental questions that must be answered even before thinking of developing and deploying an artificial intelligence models in Business.
The massive deployment of hate speech and fake news in Nigeria's digital space is too unbecoming; during the famous EndSARS protests, we saw how it led to the destruction of both public and private properties in Lagos Nigeria. There's, therefore, an urgent need to either collectively or individually rise up and wage a war against fake news. In this episode, Segun discusses how machine-learning fake news detector can be developed and deployed in the fight against fake news. While many are calling for strict regulation of social media, Segun believes that would be akin to cutting off the head of a child just because he complains of a headache. In this episode, the host also discusses that while people have the inalienable right to freedom of expression, people also have the right to privacy and protection against falsehood especially that which affects their personal life and business, and exactly how artificial intelligence maintains the balance.
Increasing sales and revenues are believed to be the single most important thing for every business person. In this episode, Segun explains how machine learning and artificial intelligence can drive innovative sales and revenues. How machine learning can model customers' purchasing behavior and patterns so perfectly to predict what other products in a store could interest customers in buying. Causing customers to buy even more outside their groceries list is perhaps one way to sell more, this episode discusses how retail business owners could leverage machine learning to achieve this and sell in a month what they sold in 3 months.
From the publisher's feed