The overselling of AI has led to a disconnect between the promised benefits and actual results, making it challenging for technology teams to integrate AI into their systems effectively.
A recent survey by BlueOptima found that even the best AI coding models succeeded less than 23% of the time when working on real production code. This is a far cry from the 85% or higher scores achieved on popular benchmarks. The study, which benchmarked 57 large language models on maintainability-oriented refactoring tasks, revealed that AI coding ROI varied dramatically by language and task, with success rates ranging from 32% in JavaScript to just 4% in C.
David Linthicum, a leading voice in technology, warns that AI is being vastly oversold, and managers need to be cautious of those who are eager to capitalize on the technology's glamour. He urges a clear-eyed, evidence-driven perspective to move past the hype and ensure that technology serves business, not the other way around.
The biggest risk with AI tools and platforms is that they may cost 10 to 20 times that of traditional systems, according to Linthicum. He also warns that decisions about organizational strategy, investment, and innovation may be based on the advice of those whose technical grasp doesn't extend beneath the surface, leading to costly overspending and strategic blunders.
When David Linthicum says AI is being vastly oversold, that means Nigerian tech professionals and startups need to be cautious of exaggerated claims and focus on developing a sharp eye for true expertise. This is crucial for navigating the high-stakes landscape of AI adoption, where companies risk resources and organizational futures if they blindly adopt AI without a clear understanding of its limits and potential.
