Based on Anthropic’s observations, this report identifies some places where the use of AI can benefit and some places where it is not helpful. For example, he said, for software developers, AI speeds up processes involved in software development, testing, documentation, and data manipulation. However, the company does not currently believe that AI is being used “meaningfully” in coordinating system installations or supervising the work of other technicians and engineers. For teachers, “we know that AI can help plan lessons and activities, but not sponsor extracurricular clubs or enforce classroom rules.”
There are limits to analysis
Extrapolating its estimates, Anthropic suggests that the current generation of AI models could increase U.S. labor productivity growth by 1.8% per year over the next decade, roughly double the rate of improvement in recent years. However, the magazine notes that “if AI changes little, these tasks could become a bottleneck and potentially act as a constraint on growth.”
We also observe that the results are based solely on conversations with Claude and therefore do not reflect the full scope of the use of AI, and that historically the greatest productivity gains in organizations have been achieved by restructuring business operations to incorporate new technology.
