AI agents do more of the work in model development, but humans still make the decisions
55% of proposals came from AI. Humans made 85% of decisions. What autonomous really means in model development.
Why it matters
A real deployment analysis reveals the gap between agent activity and actual autonomy. One-third of tasks wouldn't have been attempted without AI assistance, but humans retained decision authority across the board—a data point for practitioners evaluating agent-augmented workflows.
The key facts
11 to know769 task logs analyzed from model development project
AI agents supplied 55% of method proposals
Humans made >85% of final decisions
One-third of tasks wouldn't have been attempted without AI
Authors caution: more agent activity ≠ more autonomy
Source: research team internal analysis (not vendor marketing)
769 task logs analyzed from internal AI model development
AI agents supplied up to 55% of method proposals
Humans made more than 85% of final decisions
A third of tasks wouldn't have been attempted without AI agent input
Study warns: more agent activity does not equal more autonomy
Go to the source
The Decoderthe-decoder.com
Publisher excerpt: A research team analyzed 769 task logs from building its own AI model. AI agents supplied up to 55 percent of method proposals, but humans made more than 85 percent of final decisions. A third of the tasks wouldn't have been attempted without AI. The authors warn that more agent activity doesn't…