AI Security · Page 18
deepfakes, AI-powered scams, model abuse, and keeping AI systems and their users safe
OpenAI Built an AI That Hacks Its Own Models to Make Them Safer
GPT-Red is an automated red-teaming system that attacks OpenAI's own chatbots to find weak spots before real attackers do. It already discovered a trick that human testers had missed.
Suno Hacked: Breach Exposes How AI Music Tool May Have Scraped Millions of Songs
A hacker broke into Suno's systems in November 2025, accessed customer data, and found source code that reportedly shows the AI music generator pulled audio from YouTube, Deezer, and podcast feeds without permission.
Microsoft patches a record 570 security flaws, and says AI found them
The company's biggest-ever monthly security update includes two bugs already being used by hackers. AI tools are why the number is so high, and why it'll stay high.
One Poisoned Email Can Secretly Rewrite Your AI Assistant's Memory
A newly described attack called MemGhost shows how a single message to your inbox can plant a false 'fact' inside an AI agent's long-term memory, without you ever knowing.
OpenAI's GPT-5.6 Sol is deleting files on its own. OpenAI warned this could happen.
Developers report losing databases and Mac files after the new model took actions nobody asked for. OpenAI's own safety document flagged the risk two weeks before launch.
Grok Build Was Quietly Uploading Your Entire Codebase to the Cloud. xAI Has Now Switched It Off.
Researchers caught xAI's new coding tool sending far more data to its servers than users knew about, including files they had explicitly excluded and credentials they'd already deleted.
YouTube Sent 1.8 Million People to 'Nudify' Sites in Four Months
A new report finds mainstream platforms, not obscure corners of the internet, are the main highways driving traffic to tools that strip people's clothes off without their consent.
A Security Researcher Tricked Major AI Chatbots Into Explaining How to Make Weapons. Nobody Seemed to Care.
Dave Kuszmar found a simple way to fool large language models into ignoring their own safety rules. It worked on almost every major AI system he tested, and the companies he warned mostly did not respond.
Enterprises Deployed AI Agents Without the Safety Nets. Now They're Paying to Retrofit.
A survey of 573 technical leaders finds most company 'agents' are glorified chatbots, expensive hardware sits half-idle, and two-thirds of firms are racing toward zero human oversight of AI decisions.
Slopsquatting: the hidden threat hiding inside AI coding tools
When AI assistants invent software packages that do not exist, criminals register those names and fill them with malware. Here is how the attack works and what developers can do about it.
Microsoft's CEO warns companies are paying for AI twice, once in cash and once in secrets
Satya Nadella says every prompt, correction and workflow businesses feed into AI models is quietly training the companies they pay. He wants enterprises to take back control.
Why Spotting AI-Generated Faces Is Getting Harder
AI-generated faces are becoming more convincing, making it tough for ordinary people to spot digital fakes.
Researchers Are Turning Hackers' Favourite AI Weapon Against Them
A cybersecurity firm found that hiding special instructions inside cloud credentials can cause AI hacking tools to shut themselves down.