Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
and JavaScript.
Companies, tech workers and researchers are in a frenzy to embed agentic AI into their workflows, locked in a self-imposed race not to fall behind. There must be a better way to make use of AI technology.
Generative artificial intelligence (GenAI) tools are challenging our understanding of plagiarism. How should we deal with plagiarism of ideas if this misbehaviour is increasingly common, and it is extremely difficult to prove when GenAI is involved? Definitions of research misconduct that specifically address the use of GenAI tools are needed.
Several frameworks from different disciplines are converging on the scientific question of what it takes for a system to not just predict, simulate or reason about the world, but to act physically and intelligently within it.
A few years ago, we introduced an article format called Reusability Reports to highlight good practices in code sharing and reporting. A renewed focus on reproducibility and transparency in code reporting seems warranted, as research output has accelerated with the widespread adoption of large language models.
We introduce a framework to analyse interpretability in deep learning, by drawing on a formal notion of model semantics from the philosophy of science. We argue that interpretability is only one aspect of a model’s semantics and illustrate our framework with examples from biomedicine.
Large language models (LLMs) include not only social stereotypes but also cognitive biases. As researchers work to identify, characterize and rectify these biases, we encourage the scientific community to recognize that, although often seen as errors, cognitive biases can also reflect functional, context-specific adaptations in reasoning.
Almost 10 years ago, AlphaGo defeated one of the world’s best professional players in the complex, ancient game of Go. It was a pivotal moment that spawned new research directions and marked the beginning of a busy decade in AI development.
Less than 2% of artificial intelligence devices authorized by the US Food and Drug Agency are prognostic, with prediction horizons ranging from minutes to several years. As the number of prognostic AI devices could increase, it is important to address the accompanying regulatory and ethical challenges.
Agentic artificial intelligence (AI) frameworks are in vogue. However, implementing such systems in scientific research workflows requires clear motivations and explanations, given the risk of wasting computational as well as human resources.
Biases in artificial intelligence models have been studied predominantly through Western lenses, overlooking South Asia’s unique contexts of caste, religion, colourism and representation. This Comment highlights region-specific biases in language and vision models and calls for fairness frameworks grounded in South Asian realities.
Questions over whether neural networks learn universal or model-specific representations framed a community event at the Cognitive Computational Neuroscience conference in August 2025, highlighting future directions on a fundamental topic in NeuroAI.
Most policy proposals aimed at managing the risks of artificial intelligence (AI)-enabled weapons rely heavily on meaningful human control or appropriate human judgment for risk mitigation. This Comment argues that there are various ways humans can exert such control over AI, and that developing a careful taxonomy of these is necessary for building actionable risk-mitigation policies for warfighting AI.
