Bogus citations will get you banned from arXiv

News

arXiv clamps down on AI citations

The preprint repository arXiv has announced a one-year posting ban for researchers whose submissions are found to contain references hallucinated by artificial intelligence. Even after this penalty period, affected researchers can’t post to arXiv unless their manuscript has already been accepted at a “reputable peer-reviewed venue”, according to computer scientist Thomas Dietterich, chair of arXiv’s computer science section. Some researchers have praised the server for taking a stand; others suggest it doesn’t go far enough to tackle ‘AI slop’ in preprints.

News

AI agents accelerate drug research

Two new systems use teams of AI agents to develop hypotheses, propose experiments and analyse data in a fraction of the time it would take humans alone. The approaches still rely on human input at various stages, but when asked to identify existing drugs that might be repurposed for different conditions, they arrived at plausible answers in a matter of hours. “The goal is to give scientists superpowers,” says Google DeepMind researcher Vivek Natarajan, who helped to develop one of the systems.

Opinion

If AI does gruntwork, how do we learn?

“Much of the enthusiasm for AI tools … comes from their promise to offload work,” write anthropologist Lisa Messeri and psychologist M. J. Crockett. “But many ‘low-skilled’ tasks have conventionally been important starting points for trainee scientists.” Cleaning raw data reveals its flaws; reviewing the literature gives a holistic view; and crucial know‑how is transmitted through practical experience. Scientists should have field-wide conversations now about AI’s influence on the expertise of future colleagues, the authors argue.

Opinion

AI can’t do good science without people

With the arrival of ‘AI scientists’, it’s as well to remember that human wisdom, empathy and sheer messiness are as much part of progress as are process and efficiency, argues a Nature editorial.

Opinion

Can you mark coding assignments with AI?

Higher-education researcher Yulu Hou observed the interplay of generative artificial intelligence and human judgement in real time while watching her partner Boyan Li mark student submissions for a coding course. “In this case, AI did not remove the labour of marking, it redistributed it,” writes Hou. “It handled parts of the process, such as generating test cases and identifying possible bugs, and freed up time for the more difficult work of interpreting student thinking.”