AI, Research Integrity & Science Communication
Saturday Citations: Predicting Earthquakes; Two Types of Water; Observing Event Horizons
Article link | Chris Packham | Phys.org | June 27, 2026
Roundup highlights several recent science findings, including earthquake prediction work, unusual properties of water, and new ways of observing black-hole event horizons.
Why Have Papers by One of History's Most Prolific Scientists Been Retracted?
Article link | Science | Science | June 25, 2026
Science article examines retractions connected to a highly prolific researcher and what the case says about research integrity, peer review, and publication oversight.
As Better Chatbots Get Harder to Build, AI Turns to Other Fields for Fresh Data
Article link | Science | Science | June 25, 2026
Science article on how AI developers are looking beyond standard text data toward other scientific and technical data sources as chatbot scaling becomes more difficult.
Why Science Needs the Humanities More Than Ever
Article link | Nature Comment | Nature | June 24, 2026
Commentary argues that science needs the humanities to help interpret evidence, communicate meaning, address ethics, and respond to social division.
Medical Students Are Using a Fake Research-Paper Detector - It Doesn't Work
Article link | Frederik Joelving / Retraction Watch | Science | June 24, 2026
Article warns that tools meant to detect unreliable or fake research papers can perform poorly, creating risks for students, clinicians, and scientific trust.
'Us' Not 'Them': Scientists Must Use Their Skills to Help Stop Polarization and Division
Article link | Nature Editorial | Nature | June 24, 2026
Editorial argues that scientists can use evidence, communication, and institutional trust-building to reduce polarization rather than deepen social divides.
How AI, $1 Billion, and a Transparent Fish Could Transform Neuroscience
Article link | Zoe Beketova | Science | June 23, 2026
Science reports on an ambitious neuroscience effort using AI, major funding, and transparent fish models to study brain function at large scale.
ScienceAdviser: Did Life Begin in the Deep?
Article link | ScienceAdviser | Science | June 15, 2026
ScienceAdviser rounds up recent science stories, including origins-of-life research, new metamaterials, wireless medical devices, and other research developments.
Digital Tools Reveal Hidden Extinctions as AI Reshapes Biodiversity Science
Article link | Phys.org | Phys.org | June 15, 2026
Researchers warn that AI and digital biodiversity tools can reveal overlooked plant and fungal extinctions, but conservation still depends on reliable data about where species live and how they are changing.
Human Migration Has Surged Since 2000 - These Maps Reveal Where People Are Going
Article link | Miryam Naddaf | Nature | June 10, 2026
Nature reports that AI-assisted modeling has filled gaps in global migration data, revealing detailed population movements from 1990 to 2023.
AI Helps Reveal Large-Scale Quantum Effects Hidden in Experimental Data
Article link | Phys.org | Phys.org | June 10, 2026
Researchers used artificial intelligence to detect large-scale quantum effects that may be difficult to identify with conventional analysis of complex experimental data.
How AI Is Reshaping Discovery in Maths and Physics
Article link | Mikhail Burtsev, Yang-Hui He, Evgeny Sobko, Ananyo Bhattacharya and colleagues | Nature | June 8, 2026
Nature commentary arguing that AI is not replacing human intuition in mathematics and theoretical physics, but is changing how questions are explored, tested, and understood.
AI Is Taking on Antibiotic Resistance - Here's How
Article link | Jyoti Madhusoodanan | Nature | June 8, 2026
Nature feature on artificial-intelligence tools being used to speed antibiotic discovery and help address antibiotic resistance.
Physics-Trained Digital 'Super-Brain' Speeds Nanophotonic Design
Article link | Robert Egan | Phys.org | June 4, 2026
Article on a physics-informed AI system that speeds nanophotonic material design by reducing simulation time and requiring less training data.
The Future of Science Communication Is Not an Article Like This
Article link | Nature Editorial | Nature | June 2, 2026
Editorial arguing that science communication is changing quickly because of social media, video platforms, AI-generated content, and shifting public attention patterns.
Why AI Rules in Science Matter Now: Nature Backs Wider Debate
Article link | Phys.org / Nature | Phys.org | June 2026
Article discusses why rules for AI use in science matter now, emphasizing human judgment, transparency, fairness, and broader debate about responsible research practices.
The Last Astronomers
Article link | Joshua Sokol | Science | June 2026
Science explores how AI advances are changing astrophysics, raising questions about whether researchers will still directly interpret the universe or increasingly rely on machine systems.
Quantum Circuits Help AI Overcome Memory Limitations With Minimal New Parameters
Article link | Phys.org | Phys.org | June 2026
Quantum-computing and AI article describing how quantum-circuit approaches may help machine-learning systems handle memory limitations more efficiently.
Polymarket vs Science: Why Researchers Are Sceptical of the Prediction-Market Hype
Article link | Jenna Ahart | Nature | June 2026
Article examines whether prediction markets can reliably forecast science-related outcomes and why researchers remain cautious about using them as scientific tools.
Megalibraries Could Reshape AI-Driven Materials Discovery Faster Than Self-Driving Labs
Article link | Amanda Morris / Northwestern University | Phys.org | May 25, 2026
Materials-science article about "megalibraries" that rapidly screen huge numbers of material candidates and generate datasets for AI-assisted discovery.
AI Makes a Major Breakthrough in a Math Problem That Had Stumped Humans
Article link | Phys.org | Phys.org | May 22, 2026
Article about AI-assisted progress on a difficult mathematical problem, raising questions about machine contributions to proof and discovery.
Why AI Cannot Do Good Science Without Humans
Article link | Nature Editorial | Nature | May 19, 2026
Editorial warning that AI tools may support research, but cannot replace human judgment, interpretation, accountability, and scientific creativity.
AI Might Jeopardize the Uncertainty Required in Science
Article link | Sarah Mikula | Nature | May 19, 2026
Commentary arguing that scientific uncertainty is a productive part of research, and that AI systems could flatten or distort that uncertainty if used uncritically.
Open and Reproducible Science Without "Crisis" Framing
Article link | Sjors Scheres | arXiv | May 10, 2026
Post links to an arXiv paper while arguing that openness about scientific problems helps solve them, but "crisis" framing can be misused to attack science.
Scientists Make AI Play Battleship to Help It Do Science Better
Article link | Jennifer Ouellette | Scientific American | May 8, 2026
Article about using the game Battleship as a testing ground for improving how AI systems reason, explore, and conduct science-like problem solving.
David Attenborough Celebrates His 100th Birthday
Article link | Scientific American | Scientific American | May 8, 2026
Scientific American item celebrating David Attenborough's 100th birthday and his long role in public science and natural-history communication.
What Does "Human-Centred AI" Mean?
Article link | Neuroskeptic / Olivia Guest | PubMed | May 6, 2026
A post pointing to a paper asking what "human-centred AI" actually means, with commentary that the phrase may be more values-signaling than precise.
Scientists Invented a Fake Disease. AI Told People It Was Real
Article link | Nature | Nature | May 6, 2026
Nature article about a fictional disease being treated as real by AI systems, illustrating risks of hallucination and misinformation.
Science Stars Walk the Red Carpet: Announcing the Winners of the 2026 Breakthrough Prize
Article link | Science | Science | May 5, 2026
Science coverage of the 2026 Breakthrough Prize winners and the public celebration of major scientific achievements.
Reflection Prompts Can Slow Down Learning, Study Shows
Article link | Phys.org | May 5, 2026
Article about an AI-assisted learning experiment showing that certain reflection prompts can slow progress through programming exercises.
Opening Pandora's Box: AI Misuse to Scale Production of Human Health Research Manuscripts
Article link | Elisabeth Bik shared conference paper | arXiv | May 4, 2026
Research-integrity preprint / conference material about misuse of AI to produce large volumes of questionable health research manuscripts.
Misuse of AI by Paper Mills
Article link | Elisabeth Bik / WCRI2026 discussion | PLOS Biology reference | May 4, 2026
Conference-linked research-integrity discussion about how AI can help paper mills scale fake or low-quality research production.
Destroying Myth and Wonder
Article link | Richard Sever shared article | Journal of Cell Science | May 2, 2026
Older Journal of Cell Science article resurfaced in a discussion about scientific papers as narratives, data interpretation, and how science is communicated.
ScienceAdviser: Paging Dr. LLM
Article link | Christie Wilcox | Science | May 1, 2026
ScienceAdviser item discussing AI, medicine, synthetic biology, and recent developments from across the scientific literature.
'A Study Showed...' Isn't Enough - Scientific Knowledge Builds Incrementally
Article link | Jeffrey A. Lee / The Conversation / Phys.org | May 1, 2026
Explainer on why single studies should not be treated as final answers, and how science advances through repeated testing, revision, and synthesis.
Teams of AI Agents Boost Speed of Research
Article link | Nature | Nature | May 2026
Article on coordinated AI-agent systems that can divide up research tasks and potentially accelerate parts of the scientific workflow.
Teaching With Food Boosts Preschoolers' Science Knowledge and Vocabulary
Article link | Phys.org | May 2026
Report on research finding that food-based classroom activities can improve preschoolers' science learning and vocabulary.
Rediscovering Science: New Knowledge Hidden in Old Data
Article link | Phys.org | May 2026
Article on using AI and data science to rediscover valuable knowledge buried in old papers, graphs, tables, and experimental records.
New 'AI Scientists' Are Improving - But Reveal Their Fundamental Limits
Article link | Phys.org | May 2026
Report on AI systems designed to help with scientific work, while emphasizing the limitations that remain in reasoning, interpretation, and discovery.
Learning Physics Can Derail Some Students: New Research Helps Keep Them on Track
Article link | Phys.org | May 2026
Study coverage comparing active-learning approaches in introductory physics and astronomy courses, including SCALE-UP and investigative science learning environments.
AI 'Scientists' Promise to Accelerate Research - How Do They Work?
Article link | Nick Petrić Howe & Benjamin Thompson | Nature | May 2026
Explainer on AI systems that generate hypotheses and experimental plans, with attention to how these tools might assist or mislead researchers.
A Key Science Publishing Platform Is Cracking Down on AI Slop
Article link | The Conversation / Phys.org | May 2026
Article about arXiv and other research platforms responding to AI-generated low-quality submissions, hallucinated citations, and pressure on peer review.
AI agents set to democratise computational chemistry
Article link | The Matter Lab | Bluesky | Jan. 6, 2026 | Article: Chemistry World
Chemistry World article about AI agents and computational chemistry.
Surge in Fake Citations Uncovered by Audit of 2.5 Million Biomedical-Science Papers
Article link | Miryam Naddaf | Nature | 2026
Nature report on an audit of biomedical papers finding an increase in fabricated citations, especially since the rise of AI-assisted writing.
Network Science of Science Reviews
Article link | Tiago Peixoto | arXiv | 2026
Network-science post discussing how review systems and science itself can be studied as networks.
AI Hallucinations in Science Manuscripts
Article link | Kresten Lindorff-Larsen | bioRxiv | 2026
Preprint post warning that fabricated or "paranormal" citations in AI-assisted science writing could become a serious research-integrity problem.
