AI safety tests found to rely on 'obvious' trigger words; with easy rephrasing, models labeled 'reasonably safe' suddenly fail, with attacks succeeding up to 98% of the time. New corporate research ...

Microsoft details a new ClickFix variant abusing DNS nslookup commands to stage malware, enabling stealthy payload delivery ...

Google finds nation-state hackers abusing Gemini AI for target profiling, phishing kits, malware staging, and model ...