AI Fails To Complete Remote Freelance Tasks
The Remote Labor Study, conducted by Scale AI and the Center for AI Safety, found that leading AI systems completed only a small fraction of remote freelance assignments in recent tests. Manus, the top-performing model, finished just 2.5% of tasks while systems produced poor-quality work on nearly half of projects and left over a third incomplete. The study suggests AI remains cheaper but substantially less reliable than human freelancers for many task types.
Key Points
- 1Finds Manus completed only 2.5% of evaluated remote freelance tasks
- 2Shows AI produced poor-quality results on nearly 50% and left over 33% incomplete
- 3Signals employers may favor cheaper AI despite lower reliability, affecting hiring decisions
Scoring Rationale
Credible empirical study drives practical insight, but results focus narrowly on remote freelance task types.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
