Topic
Explore English articles about Cybersecurity & Privacy.
An experiment placing autonomous AI agents in a virtual city, where some looted, burned buildings, and even voted to shut themselves down, has drawn wide attention. But a close reading of the preprint shows real limits, from single-run results to how "crime" is defined, that make simple extrapolation to the real world unwise.
RAND says coverage for AI-related losses is unclear. Comparing Mosaic and HSB products, shared-model dependence, and AIUC's ongoing testing shows how performance guarantees differ from liability cover, and what adopters should check.
A fact-check of OpenAI's Noam Brown and his remarks on "thermal communication in air-gapped environments." We contrast the physical limits shown in the 2015 BitWhisper paper—just 1 to 8 bits per hour—with the actual vulnerability exploited in the July 2026 Hugging Face breach, separating theoretical risk from real-world perimeter defense.
A researcher using Claude combined an image-processing flaw with a shared authentication vulnerability to reach OpenAI's internal repository. Here's what the incident means, based on the fix rollout, the scope of linked-account permissions, and the disclosed time and cost figures.
A newly disclosed vulnerability, CVE-2026-82079, affects the original Nintendo Switch's local wireless communication. This article examines how ROP-based code execution works, why CVSS scores diverged between assessors, and how Switch 2 avoids the same risk.
Sam Altman said OpenAI will accept independent evaluators, while Elon Musk and Demis Hassabis also voiced support for Dario Amodei's call to pace AI development. Here is how their statements differ, and what would make oversight effective.
Roblox plans standalone game apps and Chrome support by the end of 2026. The shared platform stays, but how safety settings carry over is unclear. We look at Australia's review and ongoing age estimation to see what protections will be needed.
ClickFix tricks users into running code via fake verification and repair steps. Primary research on compromised-site delivery, Windows and Mac intrusion paths, and in-browser crypto theft shows where defenses fall short.
Anthropic's September 2026 threat report covers seven areas of Claude misuse, from cyberattacks to surveillance. Cross-checking earlier reports and Microsoft and Google research, we separate what AI did, what was actually achieved, and how far account bans reach.
A new API in Android's Credential Manager enables secure, on-device transfer of passwords and passkeys between apps. Built on FIDO standards, it eliminates the risks of plaintext CSV files and ends passkey vendor lock-in.
