The coming overreliance on AI and the superiority of human intelligence

The other day I had cause to share a thought on social media about the superiority (yet, and probably forever) of human over artificial intelligence. A few folks found the post hilarious and someone actually typed something like ‘I used to think you were smart’. Well, I lay no claim to smartness, but I just follow logic. Today, artificial Intelligence is everywhere more than before. It’s always been around in subtler forms though. Every time your phone suggests words for you, or your word document does autocorrect, underlines an error in your typing or grammatical construction, that is already artificial intelligence at work. Anytime a computer, phone, or other devices act as if they were human, that’s it. Techies went further to not only get machines and devices to act logically based on what they’ve been programmed to do and the data inside them, but to acquire new knowledge independently, and to process new knowledge in independent fashion devoid of human interference. That is what they call Machine Learning.

Scientists say a day is coming when everything will be connected to the internet. Everything! Not only computers, and phones, and your curtains in your house. But even trees will be connected to the internet. The shirt on your back. Your shoes. The grass on your lawn. How will this work? This will be when inanimate objects communicate with each other and the internet, taking decisions on their own – the lawn can order a lawnmower to come and trim it based on set criteria. Your shirt can refuse to be worn based on a load factor of germs on it if it’s dirty. Your fridge will order directly from the grocer based on depletion of supplies. And the grocery owner has to do nothing as robots select the order, package them, charge your account and deliver them to your house. Maybe you can purchase another robot that helps you take the groceries to your fridge and arrange them perfectly. Why do you have to do anything at all? After all, intelligent houses already exist where apps and robots wake you up, make your breakfast, turn the TV to your favourite channel, clean your house, start your car and drive you to work. There’s that joke about when Google ends up acquiring every company – including fast food companies – and when that day arrives, you can’t order a burger because the system has all your health details and will refuse you before you kill yourself from eating junk food.

A very interesting future beckons. And I see already an overreliance on AI, without a proper understanding of what exactly it does and how it does it. Take for example ChatGPT. This is an advanced chatbot, the type that springs up when you visit your bank’s website, asking you questions about how it can help you today. Only that this time, you can ask ChatGPT logical questions that are more complicated and give it other tasks such as to write up your CV or a business plan and whatnot. You can even set the number of words it should come up with, by typing in questions such as; such as ‘Write a 2,000 essay on Nigeria’s economy’. And in a matter of seconds, it delivers. There are advanced versions of this of course. And there are several brands out there, with new ones coming up every other week. Of late Gemini has been interrupting my work on my phone and laptop, suggesting ideas and often distracting me. There is also Copilot which pops up on my PowerPoint presentations and many more.

The problem I’ve begun to see, which prompted this article is how people are ready to cede the task of original thinking to these devices, bots and apps. When you ask ChatGPT to tell you how to manage the Naira for example, don’t go shouting wows when the reply pops up. What it has simply done is to scour the internet for related articles written by real humans (and maybe in time, written by robots themselves based on ideas from real humans). In fact, if you pay attention, the sources are usually referenced and the numbers of write-ups scoured are listed. In short, at least for now, artificial intelligence does not have the ability to think entirely outside the box or to discover its own Eureka moment. That realm, and serendipity, is left to humans.

Let’s take an example. The other day, the online news site Nairametrics did an article Titled ‘We asked ChatGPT to tell how to save the naira from depreciating in 2025…see the response’. In summary, these are the suggestions that AI came up with; 1. Boost FX Reserves, 2. Maximise Oil exports. 3. Increase incentives for exporters of goods and services (non-oil), 4. Streamline FX policies – meaning to close gap between official and parallel rates. 5. Attract foreign investments 6. Leverage diaspora remittance. 7. Increase benchmark interest rates to attract foreign portfolio investors, 8. Control money supply, 9. Combat inflation (again by tightening money supply and raising rates, 10. Combat fx speculation, 11. Institute fiscal discipline, 12. Subsidy removal and investment.

There is nothing in the above suggestions that is outside the box or ingenious. Indeed, nothing that the average Nigerian pundit with a bias for financial markets hasn’t spoken about on TV or written in articles in the past. Yet what we are looking for are not necessarily the old ideas we have always churned about but something shocking about which we will exclaim ‘why didn’t we think about this?!’. Look at how the Central Bank of Nigeria has been able to tackle foreign exchange volatility in recent times. The introduction of the Electronic Foreign Exchange Market System (EFEMS) module in the Bloomberg package has made a difference, especially with the transparency it provides. All Nigerian banks have always subscribed yearly to the Bloomberg module. But it took brainstorming sessions by human beings to decide on that option, and all of a sudden, the bogeyman of scarce fx supply was slayed – at least for the official market. Now, on that screen, every time dealers log on, they are able to see the total supply and total demand.

The danger is that the overreliance on AI and a misconstruing of how it works, its limitations and its strengths may limit even further the abilities for genuine thought among men. Africa (Nigeria) being still very vulnerable in this stage of development cannot afford to slow down due to this likely misstep, even though it is natural for human beings to be lazy and seek the fastest way out at all points in time. We cannot afford not to use our brains and come up with our greatest ideas because it suddenly looks easy for AI to do the job. Yes, AI tools must be maxed out for what it is good for – enhanced efficiency, relative accuracy, speed. This means that the situation is even more dangerous for countries like Nigeria that are used to importing every innovation. How do we limit the reliance on this innovation so that our people can think hard to solve their own problems? How do we bring our peoples’ attention to back to real, deep thinking that actually makes a lot of difference in the lives of nations?

Ritvik Nayak, writing in January 2025 in an article titled ‘10 Shocking Myths About AI That You Probably Believe”, lists the misconceptions that most people have about AI. Among what he pointed out are the myth that AI thinks like humans, that AI will take over the world, or that it will replace all jobs, or that it’s never wrong. As a fairly early adopter myself, I have seen a great many erroneous articles written by ChatGPT. It has got better over time, but errors still exist. Folks should therefore be wary about absolute dependence. You’ll hurt yourself if you ask AI to write important documents for you and you don’t proofread. Proofreading for factual correctness presupposes that you actually have real knowledge about the subject. I, for one, could actually tell an article written using AI. Usually, that profundity that comes with human thinking isn’t there. But indeed, it is efficient. Imagine purging out 5,000 words in barely two minutes. No human can measure up to that! And people will make a lot of money with AI. It’s a genius idea that can bring billions of people in the world up to speed. In fact, we are at that epoch where you really don’t have to know a lot of things. Just ask the right questions.

And this brings up a big problem that AI has already gifted to the academia. In a survey conducted in the US, about 95% of students stated that they used AI a lot, with 50% admitting that they have cheated using AI in the past. Innovations to catch up with this problem have not caught up yet. Academia is in a state of flummox. Turnitin is inadequate to detect what is what. So, plagiarism is hereby redefined. And a few errors have been made, where a writeup done organically has been misjudged as AI-written. How do we know if people actually qualify for the certificates they carry? Or is there any need for anyone to prove that they can actually do any work when AI is by the corner, beckoning anyone to use it? At this rate, what is the point of attending any school whatsoever? Thankfully though, there is still tons of work that AI has not mastered. I doubt that it could actually build a house as yet, but it can design one perfectly. The backup robots required for many physical work have not been made, or are very scarce. In all, the INSIGHT that humans can come up with has not been replicated by AI. As regards the problem with academia, I have suggested that perhaps more emphasis should be placed back on written exams. AI cheating is much easier with assignments, typed-out thesis and dissertations.

Back to my argument about human intelligence being superior to AI. Yes, that same one that some folks think is quite asinine. A few weeks ago, the financial markets were thrown into turmoil when a Chinese company announced the arrival of their own version of OpenAI’s ChatGPT. It is called DeepSeek. Those who have tried it say it is even better, faster, and more accurate compared with the current King of AI; ChatGPT. But embarrassingly for the Americans, the creators of DeepSeek say they built the entire resource with a mere $5 Million, as against the billions usually spent by the Americans.

When the Americans had a brief monopoly, the cost of putting something similar together was supposed to be the greatest barrier to entry. Now, all that is gone through the window. What caused that? Human intelligence, that’s what. In the face of much fewer resources but with human insight and ingenuousness, somebody – or a group of people in China – shocked the world. With the way DeepSeek was put together with 1/1000 the same resources used by the US for a similar project, I became sure that money wasn’t the biggest barrier to our country’s turnaround. Perhaps someone or a group of people here could compete with these geniuses in spite of lack of resources. Perhaps one brain, some brains, will deliver. We shall overcome.

Finally, as powerful as AI will get, we must never forget how to create real things. For example, AI could help a rookie engineering company write a fantastic profile that secures a major project. But AI will not help you in doing that real project. It is left to humans to interpret the engineering project as conceptualized by AI. So, not all jobs will be gone. Some that has to do with hands-on approach, leadership, emotional intelligence, and nuanced communication, will remain.