Anthropic study: AI could absorb 27-52% of researchers' time, median 38% across fields
emollick · x · 2026-11-07
Anthropic's economics team analyzed 207,000 Claude and Claude Code research sessions using privacy-preserving methods.
- Headline finding: with current capabilities and wide adoption, AI could take on 27-52% of the researcher time needed to produce a new study, with a median of 38% across fields.
- Field distribution: use is highest relative to researcher count in formal sciences (CS, math, statistics); health and lab-based physical sciences lag. Computer science accounts for 11.9% of Claude sessions and 32.7% of Claude Code sessions.
- Task mix: scientists mainly use Claude for data analysis (1/3 of Claude sessions, nearly half of Claude Code sessions) and writing — yet they rank idea generation and data gathering as most important. Specialized tools like Claude Science narrow the gap.
- Tasks where AI is theoretically strongest (top 20%) see 2-4x the usage share.
Related event: Anthropic Report: AI Could Handle 38% of Research Time, Up to 15x Speedup(5 posts)→
More from AGI Musings
- One hacker used 12 AI agents to attack hundreds of organizations across 11 countries — Graham_dePenros · 2026-11-07
- Podcast uses Hindu karma concept to frame shared responsibility for AI harms — krvarshney · 2026-11-07
- TBPN Interview: AI in Drug Discovery, GLP-1s, and China's Biotech Rise — jordihays · 2026-11-07
- Has anyone notable from the Pause AI movement flipped? ChrisGPT asks why — ChrisGPT · 2026-11-07
- NYT: OpenAI breakthrough challenges the Exponential Time Hypothesis, shaking a CS tenet — fortnow · 2026-11-07
- DREADDs: a new drug modality that works in humans, with a clear path to AI-driven design — anshulkundaje · 2026-11-07