Generative AI routinely blows up science findings

If you use AI chatbots to summarize long scientific texts, you should be careful. Often, tools exaggerate and generalize, as a study now shows. And if you ask them to be more accurate, you do the opposite. read more on t3n.de

Generative AI and scientific development

Written by Andrés García Higuera. The European Parliament’s Panel for the Future of Science and Technology (STOA) brought together Members of the European Parliament, the European Commission and researchers on 29 April 2025, to discuss the opportunities and challenges of the use of AI in science, at a workshop entitled ‘Generative AI and scientific development’. STOA Vice‑Chair Lina Gálvez (S&D, Spain) opened the event, and called for a fruitful…

Generative AI routinely blows up science findings

When summarizing scientific studies, large language models (LLMs) like ChatGPT and DeepSeek produce inaccurate conclusions in up to 73% of cases.

Most leading chatbots routinely exaggerate science findings

It seems so convenient: asking ChatGPT or another chatbot to summarise a text to quickly get a gist of it. But how accurate are they really?