On November 5, 2026, Anthropic published research analyzing 207,049 usage sessions of Claude and Claude Code,covering 34 scientific fields and 7,262 research tasks.

AI is most heavily used in computer science, mathematics, and statistics. Computer science alone accounts for 32.7% of research sessions on Claude Code.

Data analysis accounts for 33.7% of Claude sessions and 48.1% of Claude Code sessions. Writing and communicating results account for 32% and 17.6% respectively.

Conversely, AI is less frequently used for hypothesis formulation and data collection, even though 528 surveyed scientists rated these as the two most important activities.

Claude Science demonstrated the ability to handle more intensive tasks, with 82.7% of research sessions requiring a Ph.D. level and 73.3% exhibiting high AI autonomy.

According to the estimation model, current AI can help increase research output by a maximum of about 2 times if widely applied.

In a hypothetical scenario where AI automates all non-physical research work, computer science and statistics output could increase by 15–16 times, while biology and chemistry would only increase by 3–4 times.

About 40% of surveyed scientists predict that AI will be able to autonomously execute entire research projects by 2030.

📌 Conclusion: Anthropic’s research shows that AI is transitioning from a supporting tool to the capability of automating scientific research activities. However, the acceleration potential varies greatly across disciplines. Computer science and statistics benefit more due to the data-driven and computational nature of their work, whereas experimental research still relies heavily on laboratories. The 15–16x increase is merely a hypothetical scenario, not an actual result or a guaranteed forecast.

Share.
VIET NAM CONSULTING AND MEASUREMENT JOINT STOCK COMPANY
Contact

Email: info@vietmetric.vn
Address: No. 34, Alley 91, Tran Duy Hung Street, Yen Hoa Ward, Hanoi City

© 2026 Vietmetric
Exit mobile version