As Nigerian startups and global firms accelerate efforts to integrate Artificial Intelligence (AI) into product development, automation, and workplace systems, alongside these innovations, a recurring disclaimer appears across platforms: “ChatGPT can make mistakes. Check important info.” While often overlooked, industry experts say the caution highlights broader concerns about responsibility and risk in the use of AI.

In a statement yesterday, former Head of Products at OctosoA Limited, Ayodeji Ishola, drew attention to the disclaimer, stressing that it reflects the need for vigilance and accountability as businesses adopt AI-driven tools.

He said: “The disclaimer remains present regardless of how advanced the system becomes. Beneath the promise of speed and efficiency lies the admission that the system may still be wrong. And when it is, the responsibility doesn’t disappear into the technology. It returns to the person, the team, or the organisation that chose to rely on it.

“While this disclaimer may have less consequential effects in industries, such as entertainment or in low-volume production of any sort, whether physical or digital, it becomes far more critical in financial, legal and other regulated spaces, particularly in health technology, which is a sector I am quite familiar with. In those environments, an error is not always something that can be corrected in retrospect, because the consequence may already have reached a customer, a patient or a decision that cannot easily be reversed. In health technology, missing a mistake could cost a life, and that is a level of risk that, in my opinion, is simply too high to tolerate,” Ishola explained.

He noted that beneath the promise of speed, efficiency and superior innovation lies the underlying admission that the system may still be wrong, and that where the consequences of that error become serious, the responsibility does not disappear into the technology. It returns to the person, the team or the organisation that chose to rely on it.

“The important question is not only what AI can do, or how quickly we can deploy it, but what controls, standards and forms of human judgment we have put around it. Who checks the output when the workflow is automated? Who understands the consequence when the answer is wrong? Who owns the decision when AI has influenced it so deeply that the boundary between human judgment and machine suggestion is no longer clear?

“We may continue to automate execution, but accountability remains human, and as we race towards more capable systems, that disclaimer should not be treated as background text. It should be treated as a warning about foundations,” the tech expert further said.