AI Automation of Real Freelance Work Quadruples in Under Eight Months, New Index Reveals

AI automation of real freelance work has quadrupled in under eight months, with the top model now completing 16.1% of professional projects, up from a 2.5% baseline, according to the newly released Remote Labor Index.

• The Remote Labor Index (RLI) reveals that AI automation of real freelance work has surged dramatically, with the top model, Fable 5, now automating 16.1% of professional projects, more than quadrupling the 2.5% baseline recorded less than eight months ago.
• Human evaluators remain essential for accurate benchmarking, as an automated LLM judge wildly overestimates newer models' performance by up to 3x, though it still correctly ranks models relative to one another.
• Traditional time-horizon analysis does not apply to this diverse range of remote work tasks, as AI success rates do not consistently decline with longer human completion times, reflecting an uneven capability frontier across different job types.