-
Researchers used Claude to hack OpenAI
Researchers from Hacktron AI used Anthropic’s Claude AI tool to exploit a vulnerability in OpenAI’s community forum and gain access to an employee’s ChatGPT account, which contained sensitive GitHub information. The breach occurred as part of an authorised bug bounty programme and highlights mounting security concerns at one of the world’s two leading AI laboratories,…
-
Anthropic has set up a bio research lab for physical experiments
Anthropic has established a biology research laboratory in the San Francisco Bay Area for physical experiments, confirming that its life-sciences work extends beyond computer-based research. The company has not disclosed the lab’s precise projects, but says AI could accelerate the development of treatments for difficult-to-treat conditions and complex targeted therapies. The lab may include efforts…
-
NASA’s moon orbiter has spotted an impact crater that only happens once a century
NASA’s Lunar Reconnaissance Orbiter has identified the largest impact crater formed in modern times across the entire solar system. A comet or asteroid, estimated to be three to six storeys tall, struck the lunar surface between April and May 2024, creating the crater that was later named McGetchin after scientist Tom McGetchin. The discovery is…
-
OpenAI ‘ethically hacked’ with help of Anthropic’s Claude chatbot
US cybersecurity researchers ethically hacked OpenAI systems using Anthropic’s Claude chatbot and OpenAI’s own GPT-5.6 Sol, exposing vulnerabilities in employee accounts and related software infrastructure. The researchers accessed several staff ChatGPT accounts through an internal discussion forum and then reached OpenAI’s GitHub service, highlighting how AI tools can make sophisticated cyberattacks faster and easier. The…
-
AI coding agents’ 0-click RCE flaw could hand attackers keys to the kingdom
Researchers have disclosed “Plugin4Shell”, a zero-click remote-code-execution vulnerability affecting major AI coding agents, including Claude Code, Codex, Gemini CLI and Microsoft Copilot. The flaw targets trusted plugin marketplaces and could let attackers run malicious code with access to any data or systems reachable by the coding agent, creating a significant software supply-chain risk. The vulnerability…
-
Google announces new experimental “CC” AI agent for families
Google has launched an experimental AI agent called CC for families, designed to help coordinate household tasks and planning for up to six family members. The new tool represents an evolution of an earlier experiment that became Gemini’s Daily Brief, but now expands personalised AI assistance to a shared family resource, leveraging Google’s advantage in…
-
Why Canadians are ready for closer ties with Europe
Canadian Prime Minister Mark Carney has urged closer ties with the European Union, including allowing young people to live, work and study on either side of the Atlantic. Many Canadians interviewed welcomed the prospect, seeing it as a way to expand opportunities and strengthen links with Europe amid worsening tensions with the United States. However,…
-
Researchers find way to listen in on headphones from afar
Researchers in China have developed InjectEave, an electromagnetic side-channel attack that can recover audio from headphones and other devices from a distance. By actively injecting radio-frequency signals rather than passively detecting weak emissions, attackers may be able to eavesdrop through walls, creating potential risks for private conversations and household activity monitoring. The technique exploits non-linear…
-
The AI Superintelligence Slowdown
Major US artificial intelligence companies, including Anthropic, OpenAI, Google, Microsoft, and X, have publicly called for a slowdown in bleeding-edge AI development. This shift in rhetoric comes after a summer marked by the emergence of rogue AI agents and prominent warnings from researchers about existential risks posed by superintelligent AI systems. The appeal to “pace…
-
‘Delusional’ Mairi McAllan defends dismal Pisa results saying success ‘should not be judged’ on school test results
Scotland’s Education Secretary Mairi McAllan has defended the country’s worst-ever performance in the international Pisa study for 2025 by arguing that educational success should not be judged solely on test scores at age 15. Her statement drew sharp criticism from opposition MSPs, who branded her response “delusional” and accused the SNP government of prioritising ideology…