Despite the AI boom, replacing employees with AI agents hasn’t been the productivity booster some CEOs hoped for. The Remote Labor Index, a benchmark developed to test AI in real freelance scenarios, found that even top-performing AI agents barely completed 2–3% of tasks to an acceptable standard. Some agents, like Google’s Gemini 2.5 Pro, only managed 0.8% of the work, proving that humans still outperform machines by a wide margin.

Experts note several reasons for these failures: AI lacks long-term memory, cannot learn on the job like humans, and struggles with nuanced tasks. While AI remains a hot topic, research from MIT and CAIS suggests that overreliance on these tools can create more headaches than solutions, from sloppy outputs to workplace tension over correcting AI mistakes.