AI saves about 3% of your hours, and almost none of it reaches the money (okaneland.com)

Recent research by economists Anders Humlum and Emilie Vestergaard highlights a significant finding about AI’s impact on productivity. While AI can enhance task efficiency—reporting up to a 40% improvement in specific contexts like writing and customer support—the actual benefits in a real-world setting dwindle to about 2.8% of time saved, roughly an hour per week. Crucially, this incremental time savings rarely translates into measurable financial gains, with only 3-7% of productivity improvements reflecting in worker paychecks. This exposure reveals a gap between expectations from AI in controlled environments versus everyday applications. For the AI/ML community, this study underscores the need to reassess how productivity gains from AI are captured and utilized. As AI excels in specific tasks, it is imperative to focus on areas where these systems can be most effective, such as drafting and structured communication, while avoiding its use in tasks where it might produce errors. The key takeaway is that while AI can make workers faster, achieving substantial financial returns from these time savings requires intentional strategies to convert saved hours into increased output or efficiency, signaling that the real value of AI lies not in its capabilities alone but in how effectively it is integrated into workflows.