Latest cybersecurity reporting from selected sources.

Study Finds Standard Safety Tests Miss Most Agentic AI ThreatsResearchers from Stanford, MIT, Carnegie Mellon and others found that most production AI agents are vulnerable to attacks that unfold across multi-step actions. The study warns that memory, tool access and agent coordination create failure modes traditional chatbot safety testing cannot see.

Health Researchers Often Overlook Security of Historical DatasetsAn August 2025 ransomware attack on the University of Hawaii Cancer Center's epidemiology division has affected 1.2 million individuals, including personal information such as Social Security numbers of certain research study participants dating back more than 30 years.

Nearly 60% of Tech Students Said They'd Violate HIPAA If the Price Was RightBudding IT insiders can be corrupted into giving up protected health information of a very famous patient, say State University of New York at Buffalo researchers who also found a correlation between an interest in white hat hacking and illegal breaches.

Study Finds Weak Authentication Practices Across AI Agent ServersTools developers use to connect artificial intelligence tools with external applications and data sources typically are secured by static credentials such as API keys and personal access tokens, exposing AI agent systems to theft or misuse, research shows.

Default credentials, weak passwords, misconfigurations and a variety of other security shortcomings are exposing millions of medical devices and their data on the internet, said Soufian El Yadmani, CEO and co-founder of Modat, who shared recent research findings.

Default credentials, weak passwords, misconfigurations and a variety of other security shortcomings are exposing millions of medical devices and their data on the internet, said Soufian El Yadmani, CEO and co-founder of Modat, who shared recent research findings.

Robust Security Reviews of AI-Generated Code Required, Warns StudyGenies might grant magic wishes, but they inevitably come with consequences. Enter artificial intelligence coding assistants, which many developers say help them write code more quickly. New research shows that this productivity comes with a security cost, and must be mitigated.

Report Urges Indo-Pacific Cyber Shield Strategy Amid Increased Nation-State ThreatsA Center for a New American Security study found China and North Korea are accelerating cyberattacks, influence operations and infrastructure breaches across the Indo-Pacific, as researchers urge the U.S. to help develop a regional cyber shield, and deploy forward cyber teams.

Experts Aim to Probe How AI Models Reason, and Why It MattersAI researchers from OpenAI, Google DeepMind and Anthropic and others have urged deeper study into chain-of-thought monitoring, a technique to track how reasoning models arrive at answers. Their joint paper warns that transparency may erode if not prioritized.

AI Coding Assistants May Hinder, Not Help, Experienced DevelopersIf AI coding tools were meant to be the espresso shot of developer productivity, a recent study suggests they might actually be more like decaf - offering the aroma of efficiency without the buzz, based on findings from a nonprofit AI research group.

Researchers Examined 7 LLMs to Determine How they Dealt with Flawed CodeLarge language models trained on flawed data tend to replicate those mistakes, researchers found. "In bug-prone tasks, the likelihood of LLMs generating correct code is nearly the same as generating buggy code," found researchers from institutions including the Chinese Academy of Sciences.

CISOs at Organizations That Fell Victim Have a Different Story, 451 Research FindsAre your defenses against ransomware good enough to survive contact with the enemy? Don't be so sure. A new study from market researcher 451 Research finds that "overconfidence in security tooling remains an issue in the face of ransomware" for organizations that haven't yet fallen victim.

Government Nurtures Homegrown Talent and Hack-for-Hire Ecosystem, Research FindsChina boasts many of the world's most talented zero-day vulnerability researchers as well as a strict cybersecurity law compelling individuals to assist the state, and the government doesn't appear to shy away from using both those facts to its advantage, a new research study finds.

A study investigating the impact of ransomware attacks on hospitals and the ripple effect on nearby facilities is a call to action for policymakers to seriously address how those assaults can be better handled in the health sector, said researchers Rahi Abouk and David Powell.

Researchers Keep Prompts Under WrapsAcademics at a U.S. university found that if you feed a GPT-4 artificial intelligence agent public security advisories, it can exploit unpatched "real-world" vulnerabilities without precise technical information. Researchers said OpenAI asked them not to publish their prompts.

Russia, Ukraine, China and US All Top 'First-Ever World Cybercrime Index'A new study published by researchers from the universities of Oxford and New South Wales ranks Russia at the top of a global list of cybercrime hot spots and says Ukraine, China, the United States, Nigeria and Romania are home to a majority of global cybercriminal activity.

Study Shows Correlation Between Polite Language, Culture and LLM OutputIt pays to be nice, even to an inanimate chunk of code masquerading as a conversation partner, find Japanese researchers from Tokyo's Waseda University who investigated the performance of large language models under conditions ranging from rudeness to obsequiousness.

Study Looks at Impact of AI on Tasks Requiring Visual ProcessingArtificial intelligence may not steal our jobs just yet, but only because humans are currently cheaper to employ. Many of the human jobs that could be replaced with AI are not "economically beneficial" to automate - at least for now - say researchers at MIT.