research
Your photo could be all AI needs to clone your voice
A photo of someone’s face may be all an attacker needs to create a convincing synthetic voice. A new study from Australia’s national science agency explores this …
AI agents can leak company data through simple web searches
When a company deploys an AI agent that can search the web and access internal documents, most teams assume the agent is simply working as intended. New research shows how …
Scammers target international students by threatening their visa status
In 2025, the U.S. government revoked thousands of visas from international students, often without warning or explanation. According to a newly released study, this opened a …
Can your earbuds recognize you? Researchers are working on it
Biometric authentication has moved from fingerprints to voices to facial scans, but a team of researchers believes the next step could be inside the ear. New research explores …
Faster LLM tool routing comes with new security considerations
Large language models depend on outside tools to perform real-world tasks, but connecting them to those tools often slows them down or causes failures. A new study from the …
AI’s split personality: Solving crimes while helping conceal them
What happens when investigators and cybercriminals start using the same technology? AI is now doing both, helping law enforcement trace attacks while also being tested for its …
Most AI privacy research looks the wrong way
Most research on LLM privacy has focused on the wrong problem, according to a new paper by researchers from Carnegie Mellon University and Northeastern University. The authors …
A new approach to blockchain spam: Local reputation over global rules
Spam has long been a nuisance in blockchain networks, clogging transaction queues and driving up fees. A new research paper from Delft University of Technology introduces a …
When trusted AI connections turn hostile
Researchers have revealed a new security blind spot in how LLM applications connect to external systems. Their study shows that malicious Model Context Protocol (MCP) servers …
Humanoid robot found vulnerable to Bluetooth hack, data leaks to China
Alias Robotics has published an analysis of the Unitree G1 humanoid robot, concluding that the device can be exploited as a tool for espionage and cyber attacks. A robot that …
A safer way to break industrial systems (on purpose)
Cybersecurity teams often struggle to test defenses for industrial control systems without risking disruption. A group of researchers from Curtin University has developed a …
What if your privacy tools could learn as they go?
A new academic study proposes a way to design privacy mechanisms that can make use of prior knowledge about how data is distributed, even when that information is incomplete. …
Featured news
Resources
Don't miss
- Deepfake detection is losing ground to generative models
- Fragnesia: New Linux kernel LPE bug was spawned by Dirty Frag patch (CVE-2026-46300)
- Vector embedding security gap exposes enterprise AI pipelines
- Sandyaa: Open-source autonomous security bug hunter
- The hidden risk of non-human identities in AI adoption