{"id":80096,"date":"2026-04-21T00:09:38","date_gmt":"2026-04-21T04:09:38","guid":{"rendered":"https:\/\/world.edu\/?p=80096"},"modified":"2026-04-21T00:09:38","modified_gmt":"2026-04-21T04:09:38","slug":"when-ai-does-science-who-asks-the-big-questions","status":"publish","type":"post","link":"https:\/\/world.edu\/when-ai-does-science-who-asks-the-big-questions\/","title":{"rendered":"When AI does science, who asks the big questions?"},"content":{"rendered":"

\"\"<\/span><\/p>\n

They don’t just answer questions or write text. Language models like\u00a0<\/span>GPT<\/span><\/a>\u00a0,\u00a0<\/span>Claude<\/span><\/a>\u00a0, and\u00a0<\/span>Gemini<\/span><\/a>\u00a0execute code, analyze data, and even conduct experiments in robotic laboratories. Google has dubbed this idea ”\u00a0<\/span>co-scientist”<\/span><\/em><\/a>\u00a0: a virtual assistant capable of designing, planning, and executing complete experiments based on simple instructions in natural language.<\/span><\/p>\n

This technology is already yielding results.\u00a0<\/span>In collaboration with universities such as Stanford and Imperial College<\/span><\/a>\u00a0, the\u00a0<\/span>co-scientist<\/span><\/em>\u00a0has identified previously unknown biological mechanisms, suggested potential treatments for diseases like liver fibrosis, and automated parts of the scientific discovery process. Other projects, such as\u00a0<\/span>Future House<\/span><\/a>\u00a0, follow a similar path, taking the automation of science to a level that would have seemed like science fiction just five years ago.<\/span><\/p>\n

This revolution is accompanied by a shift in the habits of researchers themselves.\u00a0<\/span>A recent survey in <\/span>Nature<\/span><\/em><\/a>\u00a0revealed that 81% of scientists already use tools like ChatGPT at some stage of their work: from writing articles to generating hypotheses or drafting funding proposals. The integration of artificial intelligence (AI) into science is advancing at an unprecedented pace, but our critical reflection on its impact is not keeping up.<\/span><\/p>\n

Clear advantages, clear risks<\/span><\/strong><\/h4>\n

AI can help us write better, overcome language barriers, and explore complex data. But it also introduces significant risks.<\/span><\/p>\n

First, there is the problem of lost creativity.\u00a0<\/span>An analysis<\/span><\/a>\u00a0of more than 45 million articles and almost 4 million patents showed that, since the mid-20th century, the proportion of truly disruptive work has steadily declined.<\/span><\/p>\n

Science is advancing, yes, but increasingly so in small steps rather than transformative leaps. If we start using language models to write proposals or generate ideas, we are likely to reinforce this trend: trained in past research, they tend to reproduce dominant approaches and avoid anything radically new.<\/span><\/p>\n

An AI model can push Newton’s laws to their limits, but it wouldn’t invent the theory of relativity. It can write thousands of variations of an article on classical mechanics, but it wouldn’t ask whether Schr\u00f6dinger’s cat is alive or dead because it would never have invented quantum mechanics.<\/span><\/p>\n

A machine cannot imagine new ideas<\/span><\/strong><\/h4>\n

Deep innovation requires intuition, imagination, and the ability to challenge paradigms\u2014attributes that remain profoundly human today.<\/span><\/p>\n

There are also ethical risks. AI can fabricate data, exaggerate results, or propose experiments based on false premises, without the user noticing.<\/span><\/p>\n

It can even influence public opinion and scientific production on a massive scale, as happened with the sugar industry in the 1960s, when it promoted research that diverted attention from its health effects\u00a0<\/span>to blame fats<\/span><\/a>\u00a0.<\/span><\/p>\n

With tools capable of generating persuasive text on an industrial scale, manipulation could be far more effective. Furthermore, if advanced platforms become concentrated in the hands of a few companies or countries, the capacity for scientific discovery could be monopolized, generating new forms of scientific and technological inequality.<\/span><\/p>\n

What if a machine is both the author and the editor?<\/span><\/strong><\/h4>\n

An even more unsettling scenario is the simultaneous delegation of writing and evaluating proposals to language models. This is not science fiction: a recent study shows that one in five researchers\u00a0<\/span>already uses AI in peer review<\/span><\/a>\u00a0, and between 7% and 17% of reviews at scientific conferences on AI have been significantly modified with these tools.<\/span><\/p>\n

If one AI generates a proposal and another AI evaluates it, we enter a self-referential system where models reproduce their own biases and where human creativity is relegated. This could trap science in a spiral, negating the kind of transformative discovery that has characterized the great leaps in scientific history.<\/span><\/p>\n

An ethical framework to protect science<\/span><\/strong><\/h4>\n

To avoid these risks,\u00a0<\/span>we propose a series of ethical principles<\/span><\/a>\u00a0that allow the integration of major language models without compromising scientific integrity:<\/span><\/p>\n