Despite a global rush to embrace artificial intelligence (AI) as a cost-cutting and profit-boosting tool, a new study by the Massachusetts Institute of Technology (MIT) has found that the vast majority of businesses are failing to benefit from the technology.
According to the report, The GenAI Divide: State of AI in Business 2025, around 95% of organisations that invested in AI projects reported zero return on their investment, despite an estimated $30–40 billion poured into generative AI systems worldwide.
Researchers surveyed 300 AI deployments and interviewed 350 employees. Popular tools like ChatGPT and Copilot were found to enhance individual productivity but had little impact on company profit and loss (P&L). Only 5% of AI pilots were generating significant financial value, while most projects stalled due to integration challenges.
The study noted that the issue was less about model efficiency and more about difficulties in aligning AI with existing workflows, as well as workforce skill gaps. Nevertheless, many executives continued to fault the AI systems themselves.
Even major brands have pulled back. Taco Bell’s Chief Digital and Technology Officer Dane Mathews confirmed the company had slowed its rollout of AI-powered drive-through ordering after the system proved less effective than humans during peak hours. “At certain times, our teams are better placed to take orders. We recommend using voice AI selectively and stepping in as needed,” he said.
Meanwhile, Apple has also cast doubt on AI hype. In June, the company published a report titled The Illusion of Thinking, which argued that leading reasoning models—including Claude, DeepSeek-R1, and o3-mini—do not truly reason. Instead, Apple researchers said, such models rely heavily on pattern recognition and tend to break down when faced with higher complexity.
