AI cognitive debt — Artificial intelligence is rapidly changing how people work and learn, but it may also erode essential mental skills if used without caution, according to Steve Regis Tassembedo, a software engineer specializing in Big Data and AI. Speaking to Lefaso.net, Tassembedo described the phenomenon as a growing “cognitive debt”—a gradual loss of abilities as more mental tasks are delegated to machines.

Tassembedo outlined four stages of dependence on AI tools. The first is normal, everyday use. The second is developing a reflex to turn to AI for every problem, much as people once defaulted to search engines. The third stage is an inability to perform routine tasks without AI assistance. The final and most concerning stage, he said, is the uncritical acceptance of AI-generated answers as unquestionable truths. At this point, people may stop challenging or verifying information, which Tassembedo called “the heart of the problem.”

He cited research from the Massachusetts Institute of Technology (MIT) showing that users of advanced AI systems exhibited lower brain activity compared to those working without such assistance. Tassembedo warned that this effect is cumulative: over time, people may forget procedures, lose automatic skills, and weaken their ability to recognize and correct mistakes—a process known as metacognition.

However, Tassembedo emphasized that AI is not inherently harmful. Its impact depends on context and the user’s prior expertise. For example, an experienced journalist might use AI to draft an article but then revise and shape it, maintaining control over meaning. In contrast, a student relying on AI without first developing core skills may become dependent, achieving good grades without truly understanding or being able to defend their work.

Integrating AI into schools and universities is unavoidable, Tassembedo argued, but must be paired with systematic teaching: students should learn to question AI, verify sources, rephrase outputs, and continue practicing cognitive exercises. For African countries, he sees a unique opportunity: slower adoption allows them to learn from mistakes made elsewhere and establish healthier habits from the outset.

Ultimately, Tassembedo believes the goal is not to ban AI, but to use it as a tool that strengthens, rather than replaces, human thought. Achieving this requires political will, targeted training, and individual vigilance. Without these safeguards, he warned, society risks paying the price of a mounting cognitive debt.