What changed

OpenAI’s Economic Research team released a report detailing how over 1,800 scientists used ChatGPT during a summer pilot program. The study analyzed more than 580,000 scientific messages and surveyed 251 participants to understand how AI influences the research process.

What this means in practice

Scientists used ChatGPT across various stages of research, including reviewing literature, analyzing data, and developing ideas. Surveyed participants estimated that the same work, at the same quality, would have taken 1.9 times as long without AI. Additionally, 98% of those who answered the question about discovery timelines expected AI to accelerate discovery in their field.

Limitations and availability

By default, OpenAI does not use participants’ data from the ChatGPT for Academic Researchers program to train its models. Automated classifiers were used to sort messages by research activity, but the underlying messages and ChatGPT’s responses were not read by the research team. The study also highlights that while AI can help scientists complete tasks faster, it does not replace their involvement in the research process. OpenAI is expanding access to ChatGPT for Academic Researchers, offering free access to over 200,000 researchers globally.

Source facts for OpenAI study shows AI accelerates scientific discovery: Drag left or right, or use the arrow keys, to rotate the sphere of scientific objects. Surveyed participants estimated that the same work, at the same quality, would have taken 1.9 times as long without AI. Respondents could select multiple fields, so field groups overlap.
Original AINewsTrend explainer based on the cited source. This graphic summarizes source statements; it is not an independent benchmark.

Sources and further reading

  1. openai: Science at the frontier