Artificial Intelligence (AI) and Africa’s Quest for a Culture of Democracy
Introduction
Much has been written and said about the age of Artificial Intelligence. It is here. We have our own gurus also hyping it up from a former Deputy Prime Minister, who is also a full Professor of Robotics, to one of our biggest tech entrepreneurs. Our very own son of the soil, who is Vice President (VP) at Google, has also weighed in through various interviews, extolling the progress being made through AI. However, what kind of progress do we anticipate from AI? Let me state from the beginning that I know very little about technology in general and especially AI. My entry point into this discussion is to bring a non-expert voice to the subject. Furthermore, to state that the organization I work for is seen in some ways as a trailblazer in innovating technology for civic engagement and broader social and economic development. Africa is at some form of a crossroads- to deepen or undo democratisation, and on the other hand, to give up on the development project or pursue a different path. AI will play an important role in determining the path Africa will pursue. Do we know much about AI? A few of us, but the majority, do not know about the technology-based revolution that awaits us in both political and economic spheres. I do not believe that there is a grand and sinister AI conspiracy, but just that some of its consequences have not been adequately understood.
Technology-related advancements, especially to do with automation, usually come from a desire to free up time for humanity so that we can do more exciting stuff with our time, like go on holidays, spend time with loved ones, and generally improve quality of life. There is a huge financial return as well. Most of the billionaire class from the late 20th century up to today made their billions through technology-based innovations. It looks like there is more to be made in the AI space, and unfortunately, very few African-based businesses are part of the early pacesetters on this. Now, non-experts like me, when the term AI comes up, we think of ChatGPT, Meta, DeepSeek, Grok, and OpenAI.
These are freely available, and it’s what most of us have used. However, there are more AI tools than the ones listed above. The future is not going to be the same as the present. We have learnt so far that AI can be deployed in making the following: movies (even my own church used AI for a Father’s Day video presentation- I was proud but conflicted as well), producing material goods such as clothing, and assembly line production in the automotive industry. Furthermore, AI or its cousin, the algorithm, plays a very important role in how we consume information. For those of us still using Facebook, we do not choose the videos that come up in our feed, yet there is a certain consistency. Artificial Intelligence studies our online behaviors, one, two clicks they know you, and they will feed you based on what you choose. There was a time when I was getting all these shoe adverts-then I remembered that I had tried to shop online (not on Facebook) for a pair of shoes. So yes, they do not only learn from your online behaviors based on one platform. This is not unique to Facebook. It’s quite common for you to see it on YouTube and in many other places.
In other words, the machine has learnt about your interests and is doing its job aligning.
Now let’s delve into the dangers of all these things.
AI, the Truth and Escalation of Conflict
Mass communication scholars have always defended the need for free media as a basis for a strong democratic culture. How does AI deepen or undo free media? From what I have read, AI can push an agenda. Again, it depends on the algorithm. First, let me explain the algorithm in layman’s terms. It is a system used to decide what kind of information one should receive based on their previous online activities, and here is a strange one also, based on your location. This is part of an effort to ensure that one is receiving relevant news, whether on X (formerly known as Twitter), Facebook, YouTube, or even WhatsApp channels. Harari in Nexus has a fascinating case study of how Facebook feeds were used to escalate violence against the predominantly Muslim Rohingyas of Burma. A fringe group called the Arakan Rohingya Salvation Army (ARSA) staged coordinated attacks on the police in Rakhine state of Myanmar (Aljazeera, 2022). The army retaliated. It killed 400 armed fighters, but others argue that most of the dead were civilians. The violence against the Rohingyas escalated. More than 120,000 fled to Bangladesh. Without simplifying the conflict, Harari demonstrates (through detailed reports) how Facebook was used as a platform to spread hate messages against Rohingyas, like how radio was used to spread hate messages against the Tutsis of Rwanda in the early 1990s. There were a few Buddhist clerics who were calling for the extermination of Rohingyas, and their messages went viral on Facebook. Instead of the algorithm censoring this as hate message, it saw it as popular because of the engagements and number of times it was opened.
Global human rights group Amnesty International was one of the first to call out Meta, the owner of Facebook. “In 2017, the Rohingya were killed, tortured, raped, and displaced in the thousands as part of the Myanmar security forces’ campaign of ethnic cleansing. In the months and years leading up to the atrocities, Facebook’s algorithms were intensifying a storm of hatred against the Rohingya, which contributed to real-world violence,” (Agnès Callamard, Amnesty International’s Secretary, Amnesty International, 2022). There were some moderate clerics who pushed for tolerance and peace. Their messages were not popular, so the algorithm did not push those messages to many in Myanmar. In a nutshell, the algorithm has the power to ensure peace or violence. However, now the algorithm’s purpose is to maximise engagements (the number of times you stay on the page), likes, and downloads. There is limited investment in trying to make sure that the algorithm is trained to tell the truth. It is trained to enhance the profitability of the App by making sure that people remain glued to their screens, as this is what drives ad revenues. Amnesty further noted that ‘Meta profits when Facebook users stay on the platform as long as possible, by selling more targeted advertising. The display of inflammatory content, including that which advocates hatred, constituting incitement to violence, hostility, and discrimination is an effective way of keeping people on the platform longer.
Facebook is not alone. According to The Guardian (2025), Elon Musk’s Grok (the AI tool affiliated to X) in mid-May 2025 admitted that it was ‘instructed by my (its) creators to accept genocide (in South Africa) as racially motivated’. CNBC, a news channel, weighed in to state that AI chatbots such as Grok can be manipulated by human beings. A new term has emerged to explain the faults of AI: ‘hallucinations. These are moments when AI tools fabricate information entirely. Isn’t that enough to cause us to worry (Vanian ,2025)? There is no consensus yet on what causes these hallucinations. This incident was particularly alarming, as it demonstrates how those same techniques can be deliberately abused to produce misleading or ideologically motivated content (Foulds et al., 2024)
Increased Dependency on AI
We are the Google generation. We have coined ‘to Google’ as a synonym for the word search. Many articles have been written based on searching for information on the internet. One can search for articles, videos, books, and other forms of information on Google. If the information from a particular source did not come, it was never Google’s fault, but maybe the source’s failure to do something called Search Engine Optimization (SEO). We have gone through various rounds of these as an organization just to make sure that our information is available across a wide variety of search words used when looking for information. It has worked somewhat. However, that is under threat. First, Google is not the only search engine; there are many others, but it commands more than 90% of web traffic that starts from a search engine (Germain, 2025). People use Google Search five trillion times a year (Argawal and Jain, 2025). Most of the internet-based searches start from Google. In other words, as we write, Google is responsible for what you have access to on the Internet, and probably more than 25% of the knowledge you have acquired has been found through Google. As already stated, it always looked like a very innocent platform where you asked a random question like ‘what is the best city to live in’ and you would get many articles from newspapers, archives, some platforms, etc. But since a couple of years ago, Google added what it calls an AI overview. Before you get to other sources related to your search, Google would try to answer your question itself to make it easy for you to save time (I suppose). It was always a small summary, and for me, it looked like it was replacing Wikipedia, which always used to be at the top. Side note- I spent time looking at Wikipedia’s search engine strategy because I always felt that if I could be better at that, then my organization would always feature at the top.
According to Germain (2025) of the BBC, Google is about to change how search works. He cites Google’s CEO’s announcement of what he called the new end-to-end AI search experience, and he emphasized that it’s a total reimagining of search. The AI overview was always optional. One can skip it and go to other search results. AI models will be very different. According to various experts, AI modes will replace traditional search altogether; instead, a chatbot will effectually create a miniature article to answer your question. You will lose your ability to carry out more searches and rely on what the AI tool gives you as the most adequate answer. There are concerns that many businesses that rely on Google search will crash. There is also a concern that the truth may be sacrificed. New websites will have to optimize search for the AI chatbot, which will determine what is included or excluded from the answer. Could the era of the open web be overtaken by an all-knowing chatbot AI that determines what we find when we search? How will that affect democratic discourse?
The second argument focuses on the possibility of a decline in our problem-solving skills as humans. The AI Index Report (2025) posits that, in 2023, researchers observed a dramatic leap in AI performance within just one year. Notably, language model agents demonstrated the ability to outperform humans in specific programming tasks, particularly when working under strict time constraints. This signals a growing efficiency and capability in AI’s problem-solving and technical execution compared to its human counterparts. We have entered an era where machines are getting smarter than humans. For the record, man has been beaten by a machine at chess and at go, a Chinese game. Let me explain a bit, ‘go’ is one of the oldest and most sophisticated board games in the world, originating in China over 2,500 years ago. It is a strategic game for two players where the goal is to surround more territory than the opponent. It is considered one of the four essential arts of ancient Chinese scholars. It represents strategy balance and discipline, often compared to military tactics. Why does ‘go’ matter? It is strongly associated with the meritocracy in China. Yet in October 2015, Google’s DeepMind won against the European Go champion Fan Hui 5-0 (Sneed, 2016). In Chess, machines have won several matches against world champions. According to Cystal Springs Uplands School (2014), it started way back in 1997 when IBM’s Deep Blue did something that no machine had done before. It became the first computer system to defeat a reigning world champion. What is worrying is this, whereas Kasparov’s 1996 match was against specifically designed hardware that took millions of research and development. Today, even low-end mobile devices can defeat the top human. (Side note, I have been defeated several times by a machine, but I am also not that much of a top player). If these two games are the true test of our intelligence, then we are in deep trouble.
We have created something with the potential of becoming smarter than us, and yet we still believe we can control it.
Again, I am not suggesting any conspiracy theories, but if these machines are smart because they have capabilities to learn, and we as humans are devoting more attention to leisure and let the machines fix the world’s problems for us, what will be the eventual outcome?
Are we getting smarter?
It is said that AI has the potential to, or is already, revolutionizing various aspects of modern life. Various AI tools, as described above, comprising virtual assistants and recommendation algorithms to complex decision support systems, are meant to improve the quality of lives. But are they making us smarter or dumber? The 2% working in AI-related technology are getting smarter. How about the rest of society? Are our cognitive skills developing and responding to challenges? The increasing reliance on AI for the retrieval of information and decision-making raises questions about how these technologies influence users’ critical thinking abilities.Gerlich (2025) argues that “…growing evidence demonstrates that overreliance on these tools can lead to cognitive overoffloading.” This occurs when individuals delegate cognitive tasks to external aids, reducing their engagement in deep, reflective thinking. Gong and Yang (2024) put forward that this may lead to an overall decline in cognitive engagement and skills development. The world has always, up to now, been driven by the humans’ quest for better, and as we fixed one set of problems, we moved onto the next—in the process, our cognitive skills kept on improving until now, when we are delegating problem tasks to tools.
Even before AI Universities were struggling with increased cases of plagiarism. Many essays were based on cutting and pasting from the internet. AI makes it even worse. Students can ask ChatGPT, Meta, or DeepSeek to create an essay with references, and all this can be done in less than fifteen minutes. Tools like Turnitin are struggling to differentiate between submissions that have been 100% developed by humans or by AI. Adams (2024) conducted a survey among 157 Universities in the UK, and 130 responded. Findings revealed that between 2023 and 2024, there were 7,000 cases of students caught using AI to cheat, 5.1 for every 1,000 students. The previous year, there were just 1.6 cases per 1,000 students. Most of the AI-generated content goes undetected. Findings further highlighted that the University of Reading conducted an experiment and submitted AI-generated answers through fake student profiles to its professors, but it was not detected. All this poses serious concerns about academic integrity, and considering the scale, it will only get worse. Some universities on the continent require students to submit printed assignments, meaning the marker will simply review the document and, given the sheer number of scholars, may not have the capacity to cross-check all submissions thoroughly. Besides cheating, the AI era has the potential of reducing diversity of thought, which is a crucial aspect to building healthy democracies. The diversity of information available on any subject is under threat thanks to algorithms and back-to-back AI searches. Unpopular views risk being eliminated from public discourse. How will the algorithm be governed to ensure that all available information is made available?
Conclusion
It is important to consider the environment or context in which AI is evolving. We are living through a moment of value-based polarization and increasing intolerance toward oppos=ing views. Can AI be trusted to judiciously manage diverse perspectives and allow users to form judgments based on the information available? I have already referred to X’s Grok, which admitted to being programmed with the assertion that genocide is occurring in South Africa. What might happen during cliffhanger elections? Would the owners of an algorithm be tempted to interfere to ensure that one side prevails? How will controversial subjects be discussed in public spaces, such as the internet and chat forums? Will AI seek to influence or tilt the balance of discourse? Let’s add another dimension: how will these profit-driven AI tools reconcile the pursuit of financial returns with the responsibility to provide balanced and impartial information? Have we capitulated too soon to a new form of existence where machines begin to assume godlike influence? There is a real danger that these seemingly ordinary tools of convenience can be manipulated for ideological purposes. According to BBC News (2014), Stephen Hawking issued a sobering warning: “A super intelligent AI will be extremely good at accomplishing its goals, and if those goals are not aligned with ours, we are in trouble.” Have we already arrived at that moment of trouble?
Be on the lookout for Part 2 of this discussion focusing on AI and Africa’s Quest for Economic Transformation
Gong, C., & Yang, Y. (2024). Google effects memory: a meta-analytical review of the media effects of intensive Internet search behavior. Frontiers in Public Health, 12.https://doi.org/10.3389/fpubh.2024.1332030
Gerlich, M. (2025). AI tools in society: Impacts on cognitive offloading and the future of critical thinking. Societies, 15(1), 6.https://www.mdpi.com/2075-4698/15/1/6
