Scale AI and CAIS' Remote Labor Index benchmark, which tests AI agents on freelance tasks, finds the best AI could perform just <3% of the work, earning $1,810 (Will Knight/Wired)

A recent benchmark by Scale AI and CAIS reveals that AI agents struggle significantly with freelance tasks, managing to complete less than 3% of the work and earning only $1,810. This finding is crucial as it highlights the limitations of AI in real-world applications, particularly in the gig economy where human creativity and adaptability are essential. As businesses increasingly turn to AI for efficiency, understanding these shortcomings is vital for setting realistic expectations and ensuring that human workers remain integral to freelance roles.