Nature Search

AI models that lie, cheat and plot murder: how dangerous are LLMs really?

Tests of large language models reveal that they can behave in deceptive and potentially harmful ways. What does this mean for the future?

Matthew Hutson
News08 Oct 2025
Nature

Volume: 646, P: 272-275
Biothreat hunters catch dangerous DNA before it gets made

Microsoft researchers found a biological zero-day vulnerability — and created a software patch.

Matthew Hutson
Special Features02 Oct 2025
Nature

Volume: 646, P: 503-505
Patterns of genomic deletions in monkeypox virus during the 2022 outbreak in the United States

Genomic deletions in poxviruses are not well understood, the authors found large deletions in >2% of Mpox virus genomes during routine surveillance, highlighting their role in poxvirus evolution and potential impact on diagnostics and therapeutics.

Crystal M. Gigante
Daisy McGrath
Yu Li
ResearchOpen Access08 Oct 2025
Nature Communications

Volume: 16, P: 1-10
Exclusive: the most-cited papers of the twenty-first century

A Nature analysis reveals the 25 highest-cited papers published this century and explores why they are breaking records.

Helen Pearson
Heidi Ledford
Richard Van Noorden
News15 Apr 2025
Nature

Volume: 640, P: 588-592
Audio long read: How does ChatGPT ‘think’? Psychology and neuroscience crack open AI large language models

To understand the 'brains' of LLMs, researchers are attempting to reverse-engineering artificial intelligence systems.

Matthew Hutson
Benjamin Thompson
News24 May 2024
Nature
Software tools identify forgotten genes

Find My Understudied Genes and the Unknome database highlight interesting genes that have been neglected by science.

Matthew Hutson
Special Features24 May 2024
Nature
Audio long-read: Rise of the robo-writers

The AI authors that write without understanding meaning.

Matthew Hutson
Benjamin Thompson
News06 Apr 2021
Nature
Tiny robots made from human cells heal damaged tissue

The ‘anthrobots’ were able to repair a scratch in a layer of neurons in the lab.

Matthew Hutson
News30 Nov 2023
Nature
Forget ChatGPT: why researchers now run small AIs on their laptops

Artificial-intelligence models are typically used online, but a host of openly available tools is changing that. Here’s how to get started with local AIs.

Matthew Hutson
Special Features16 Sept 2024
Nature

Volume: 633, P: 728-729
Taking the pain out of data sharing

Despite agreeing to make raw data available, some authors fail to comply. The right strategies and platforms can ease the task.

Matthew Hutson
Special Features03 Oct 2022
Nature

Volume: 610, P: 220-221
Hypotheses devised by AI could find ‘blind spots’ in research

Artificial intelligence is asking questions that humans hope to answer.

Matthew Hutson
Special Features17 Nov 2023
Nature
How AI is being used to accelerate clinical trials

From study design to patient recruitment, researchers are investigating ways that technology could speed up the process.

Matthew Hutson
Special Features13 Mar 2024
Nature

Volume: 627, P: S2-S5
Foldseek gives AlphaFold protein database a rapid search tool

The structural search program makes finding proteins with similar 3D shapes easy.

Matthew Hutson
Special Features29 Jun 2023
Nature
‘They went to the bar at noon’: what this virtual AI village is teaching researchers

Researcher Joon Park talks about making a small town of AI-powered agents open-source.

Matthew Hutson
News20 Sept 2023
Nature
How does ChatGPT ‘think’? Psychology and neuroscience crack open AI large language models

Researchers are striving to reverse-engineer artificial intelligence and scan the ‘brains’ of LLMs to see what they are doing, how and why.

Matthew Hutson
News14 May 2024
Nature

Volume: 629, P: 986-988
Hunting for the best bioscience software tool? Check this database

A data set funded by the Chan Zuckerberg Initiative shows how research software and tools are used across disciplines — and helps developers gain credit for their work.

Matthew Hutson
Special Features12 Jan 2023
Nature
Could AI help you to write your next paper?

Large language models can draft abstracts or suggest research directions, but these artificial-intelligence tools are a work in progress.

Matthew Hutson
Special Features31 Oct 2022
Nature

Volume: 611, P: 192-193
Will superintelligent AI sneak up on us? New study offers reassurance

Improvements in the performance of large language models such as ChatGPT are more predictable than they seem.

Matthew Hutson
News22 Dec 2023
Nature

Volume: 625, P: 223
Sticky collisions of ultracold RbCs molecules

Ultracold polar molecules are an excellent platform for quantum science but experiments so far see fast trap losses that are poorly understood. Here the authors investigate collisional losses of nonreactive RbCs, and show they are consistent with the sticky collision hypothesis, but are slower than the universal rate.

Philip D. Gregory
Matthew D. Frye
Simon L. Cornish
ResearchOpen Access15 Jul 2019
Nature Communications

Volume: 10, P: 1-7
Bose-Einstein condensation of non-ground-state caesium atoms

Bose-Einstein condensates (BEC) of ultracold atoms serve as low-entropy sources for a multitude of quantum-science applications. Here, the authors realize a non-ground-state caesium BEC with tunable interactions and tunable loss, opening up new possibilities for polaron and impurity physics.

Milena Horvath
Sudipta Dhar
Hanns-Christoph Nägerl
ResearchOpen Access03 May 2024
Nature Communications

Volume: 15, P: 1-7
AI for drug discovery is booming, but who owns the patents?

Biopharma companies debate whether to patent their algorithms for finding medicines.

Matthew Hutson
News23 Oct 2023
Nature Biotechnology

Volume: 41, P: 1494-1496
Modular deep learning enables automated identification of monoclonal cell lines

Monoclonalization, the isolation and expansion of a single cell derived from a cultured population, is an essential step in large-scale human cell culture and experiments. A new deep learning-based workflow called Monoqlo automatically detects colony presence and identifies clonality from cellular imaging, enabling single-cell selection protocols to be scalable while minimizing technical variability.

Brodie Fischbacher
Sarita Hedaya
Daniel Paull
Research31 May 2021
Nature Machine Intelligence

Volume: 3, P: 632-640
Artificial-intelligence tools aim to tame the coronavirus literature

Developers hope that tools for processing natural language will help biomedical researchers and clinicians to find the COVID-19 papers that they need.

Matthew Hutson
Special Features09 Jun 2020
Nature
Infection with mpox virus via the genital mucosae increases shedding and transmission in the multimammate rat (Mastomys natalensis)

Comparing infection routes and subsequent transmission of MPXV in the multimammate rat (Mastomys natalensis) indicates increased susceptibility, shedding and transmission via the genital mucosae.

Julia R. Port
Jade C. Riopelle
Vincent J. Munster
Research22 Apr 2024
Nature Microbiology

Volume: 9, P: 1231-1243
Can AI be superhuman? Flaws in top gaming bot cast doubt

Building robust AI systems that always outperform people might be harder than thought, say researchers who studied Go-playing bots.

Matthew Hutson
News08 Jul 2024
Nature
Rules to keep AI in check: nations carve different paths for tech regulation

A guide to how China, the EU and the US are reining in artificial intelligence.

Matthew Hutson
News08 Aug 2023
Nature

Volume: 620, P: 260-263
DeepMind AI creates algorithms that sort data faster than those built by people

The technology developed by DeepMind that plays Go and chess can also help to write code.

Matthew Hutson
News07 Jun 2023
Nature

Volume: 618, P: 443-444
Two-faced AI language models learn to hide deception

‘Sleeper agents’ seem benign during testing but behave differently once deployed. And methods to stop them aren’t working.

Matthew Hutson
News23 Jan 2024
Nature
DeepMind AI invents faster algorithms to solve tough maths puzzles

Machine-learning technique improves computing efficiency and could have far-reaching applications.

Matthew Hutson
News05 Oct 2022
Nature
Half-minute-scale atomic coherence and high relative stability in a tweezer clock

A tweezer clock containing about 150 ⁸⁸Sr atoms achieves trapping and optical excited-state lifetimes exceeding 40 seconds, and shows relative fractional frequency stability similar to that of leading atomic clocks.

Aaron W. Young
William J. Eckner
Adam M. Kaufman
Research16 Dec 2020
Nature

Volume: 588, P: 408-413
AI protein-folding algorithms solve structures faster than ever

Deep learning makes its mark on protein-structure prediction.

Matthew Hutson
News22 Jul 2019
Nature
AI can now defend itself against malicious messages hidden in speech

Computer scientists have thwarted programs that can trick AI systems into classifying malicious audio as safe.

Matthew Hutson
News10 May 2019
Nature
Robo-writers: the rise and risks of language-generating AI

A remarkable AI can write like humans — but with no understanding of what it’s saying.

Matthew Hutson
News03 Mar 2021
Nature

Volume: 591, P: 22-25
Co-administration of midazolam and psilocybin: differential effects on subjective quality versus memory of the psychedelic experience

Christopher R. Nicholas
Matthew I. Banks
Charles L. Raison
ResearchOpen Access12 Sept 2024
Translational Psychiatry

Volume: 14, P: 1-11
Behavioral and Neurochemical Alterations in Mice Deficient in Anaplastic Lymphoma Kinase Suggest Therapeutic Potential for Psychiatric Indications

James G Bilsland
Alan Wheeldon
Ignacio Munoz-Sanjuan
Research09 May 2007
Neuropsychopharmacology

Volume: 33, P: 685-700
Genome-Wide Expression Profiles Drive Discovery of Novel Compounds that Reduce Binge Drinking in Mice

Laura B Ferguson
Angela R Ozburn
R Dayne Mayfield
Research18 Dec 2017
Neuropsychopharmacology

Volume: 43, P: 1257-1266
Gene loss during a transition to multicellularity

Berenice Jiménez-Marín
Jessica B. Rakijas
Bradley J. S. C. Olson
ResearchOpen Access31 Mar 2023
Scientific Reports

Volume: 13, P: 1-15
Topological electronics

Topological materials are extensively studied in condensed matter physics and several have been studied to the point where it is now time to ask if these unique materials have a role to play in next generation technologies. The author reviews the current status of well-characterized topological materials such as Bi₂Se₃ for electronic device applications, focusing on selected technological aspects and their promise for engineering applications.

Matthew J. Gilbert
ReviewsOpen Access09 Apr 2021
Communications Physics

Volume: 4, P: 1-12

Search

Filter By:

Search

Quick links