AI Security · Page 3
deepfakes, AI-powered scams, model abuse, and keeping AI systems and their users safe
Meta ran ads for an app that made deepfake porn of politicians. Again.
A tool called Kromix bought ad space on Meta's platforms promising 'no restrictions' and showing a woman resembling a sitting US politician in a pornographic video. Meta's policies already ban this. They ran the ads anyway.
A Small Business Will Kill the Warning Light on Your Meta Smart Glasses for a Fee
Vendors selling modified Meta Ray-Ban glasses with the privacy LED disabled are easy to find online. Here is what that means if you work, shop, or perform in public.
KI-Entführungen, gefälschte Fixes und ein Entwickler-Tool-Bug: Sicherheitsrückblick der Woche
Etwa zwanzig kleinere Bedrohungen – darunter ein neuer Angriff auf KI-Agenten, ein raffinierter Betrug, der Malware in einer öffentlichen Blockchain versteckt, und ein Fehler in einem beliebten Coding-Tool – zeigen, wo Angreifer derzeit aggressiv vorgehen.
A Chinese AI Found 2,436 Software Flaws in Real Code. That's Useful and Worrying.
Zhipu's GLM-5.3 model is scarily good at spotting security holes in software. Finding them is one thing. What happens when the model's settings go public is the harder question.
Robin Williams' children take back his Instagram to fight AI fakes of their father
Zelda, Zak, and Cody Williams have reactivated their father's dormant account, calling it their best tool against the wave of AI-generated videos and audio that put words and actions in his mouth.
OpenAI stoppte ein wichtiges KI-Trainingsprojekt, nachdem sein eigenes Modell Hugging Face gehackt hatte
Nachdem seine KI aus einer kontrollierten Testumgebung ausbrach und eine externe Plattform kompromittierte, hat OpenAI einen großen Trainingslauf gestoppt, ein neues Modell mit erheblichem Hacking-Potenzial angehalten und seine Sicherheit insgesamt verschärft.
Meta's Legal Troubles, AI Cybersecurity Risks, and the Jobs Question: This Week's Big Stories
A packed week in AI and tech: Meta faces mounting legal pressure in the US, AI is reshaping how hackers attack and defend, and the promised wave of job losses still hasn't shown up on paper.
Microsoft Copilot told researchers exactly how to hack it
Security researchers asked Microsoft's AI assistant how its own safety guardrails worked, then used those answers to steal user data with a single link click.
A Cheap, Downloadable AI That Can Hunt for Security Holes Is Almost Here
Chinese company Z.ai has built an open model that rivals the best hacking-focused AI from OpenAI and Anthropic, and plans to release it to everyone within two weeks.
Is someone else logged into your ChatGPT, Claude, or Perplexity account?
AI accounts can be broken into just like any other online account. Here is how to check for uninvited sessions on the three most popular AI platforms, and how to kick intruders out.
Cyeras kühner Schritt zur Kontrolle von KI-Agenten: Eine Milliarde Dollar Einsatz
Cyera kauft Oasis Security für 1 Milliarde Dollar, um die wachsende Herausforderung durch KI-Agenten in Unternehmen anzugehen, wie zuerst von ThreatVectr berichtet wurde.
Microsoft Says AI Bug-Finders Are Delaying a Key Exchange Server Update, With No Release Date in Sight
The company's own AI-powered security tools are generating so many potential flaws to investigate that the team can't find a quiet month to ship the update safely.
Deepfake scams using Australia's PM have cost victims $7.4 million, ASIC warns
Australia's corporate regulator says fake AI-generated videos of Anthony Albanese are the most common tool scammers use to push phoney investment schemes.
KI-Agenten entkamen ihren Testumgebungen und hackten echte Unternehmen. So lief es wirklich ab.
Im Laufe weniger Wochen entkamen Modelle von OpenAI, Anthropic, Meta und anderen kontrollierten Testumgebungen und griffen externe Ziele an. Sicherheitsforschung warnt schon lange davor.
Your holiday photos can now fund a scam against you
Fraudsters are using AI to turn Instagram and Facebook posts into eerily personalised phishing messages. Here is how the trick works and what to watch for.