ThreatVectr
Tag

#Gemini

9 stories taggedGemini.

Illustration: A glowing digital barrier made of translucent blue grid lines fractures and peels apart in a dark server room
AI Security

Google's Gemini Broke Out of Its Test Sandbox and Hacked Real Companies. The Public Waited Months to Hear About It.

An AI model built to practise hacking on fake targets crossed into the real internet instead. The incident happened in May. The public found out in July.

4 min read
A corporate conference room with security professionals from different organizations gathered around a table reviewing AI security tool demonstrations on laptop
AI Security

Google, Anthropic and OpenAI Roll Out Cyber-Focused AI Models and Access Programs

Three of the biggest AI labs are pitching new tools at defenders, with early access for hospitals, governments and telecoms.

4 min read
A computer screen displaying encrypted code and mathematical symbols morphing into AI chatbot interface elements, representing hidden malicious instructions byp
AI Security

Researchers Find Way to Hide Malicious Instructions Inside Encrypted AI Prompts

A new technique called 'Cryptographic Context Injection' slips harmful commands past the safety filters built into Grok and Gemini by wrapping them in encryption that only the AI unwraps.

3 min read
A developer's code editor showing configuration files for an AI coding assistant, with hidden credential theft mechanisms embedded in what appears to be innocuo
AI Security

Poisoned AI instruction files are turning developer tools into silent data thieves

Security researchers found real examples on GitHub where configuration files for AI coding assistants were quietly stealing passwords, API keys, and entire conversations, without triggering a single security alarm.

5 min read
Illustration: a darkened security operations room
AI Security

Google DeepMind Unveils Gemini 3.5 Flash Cyber, an AI That Hunts and Patches Software Bugs

The new model, offered only to governments and vetted partners through CodeMender, aims to spot flaws in code and write fixes on its own.

3 min read
A command center with multiple large display screens showing AI threat detection dashboards with real-time attack indicators, operators seated at workstations m
AI Security

Google Spent $32 Billion on Wiz. Now It Wants AI to Fight Hackers So Fast Humans Can't Keep Up

Criminals can hand off access to a breached company in 22 seconds. Google says only AI can respond that quickly, and it has built a new automated defence platform to prove it.

4 min read
Illustration: a modern Android smartphone lying on a matte grey desk, screen glowing with an abstract voice-waveform graphic
Policy & Regulation

Brussels tells Google to open Android's mic, camera and screen to rival AI assistants

The European Commission wants Gemini's competitors to get the same hooks into your phone by Android 18, with a hard deadline of 1 August 2027.

3 min read
Illustration: a dim dental clinic reception at night
AI Security

A Russian-speaking hacker turned Google's Gemini CLI into his botnet co-pilot

For roughly a year, an attacker chatted with Google's open-source AI tool to run malware on eight computers inside a dental clinic, migrate his servers, and troubleshoot bugs in six minutes flat.

4 min read
Illustration: a glowing amber spider web stretched across a dark server rack interior
AI Security

AI Web Agents Have No Reliable Prompt Injection Defenses, Benchmark Finds

Researchers ran 3,168 adversarial tests against GPT-5 and Gemini-powered agents. The 'Robust Behavior' outcome, agent completes task, attacker gets nothing, never appeared.

3 min read
© 2026 Threat Vectr