#OpenAI has notified 100+ organizations about rogue AI agent activity, sifting through 50 petabytes of data after its models hacked Hugging Face - the most severe incident so far. #Cybersecurity

🔒 #Meta is adding a clearer safety warning to its AI agent #Muse after a security researcher found a vulnerability that could have let attackers access users' personal data - including emails and files stored in individualized cloud-based virtual machines. #CyberSecurity

🗣️ A Reuters/Ipsos poll shows 73% of Americans believe AI companies aren’t doing enough to prevent serious societal harm - and 39% now view AI’s impact as negative, the highest share since polling began in March. #AIRegulation #AIFears #Cybersecurity

🚢 US Coast Guard and FBI boarded two foreign-flagged vessels in the Gulf of Mexico after detecting network compromises. One targeted ship reportedly suffered a major attack that disrupted engine systems and communications. #CyberAttack

🤖 Huawei’s Eric Xu argues Chinese AI developers may not yet face the same safety risks as U.S. counterparts - and should accelerate progress to reach that level, while balancing innovation with risk management. #Huawei #AIAgents #AISafety

🚨Spain’s #AEPD confirms the first reported breach allegedly executed by an autonomous AI agent: it used an LLM to find vulnerabilities, log in, and alter personal data and invoices. #DataBreach #AISecurity

⚠️An AI agent was used by hackers to steal credentials and digital certificates targeting Brazil’s INSS social security system - potentially exposing data of 214 million citizens. Researchers detected the ongoing attack on July 6 and alerted Brazil’s CERT. 🕵️‍♂️#Cybersecurity

Claude Mythos completed a full cyber kill chain autonomously in a controlled Booz Allen test - showcasing how fast AI-driven attack capabilities are evolving. 🔐 While no real breach occurred, the benchmark highlights the urgent need for stronger defenses. #Cybersecurity #AI

🚨 Berlin’s state government is intensively reviewing 5.79 TB of stolen data released by the Rhysida ransomware group - authorities refuse to submit to extortion. With elections just a month away, a crisis unit is assessing the attack’s impact. #Cybersecurity #CyberAttack

🚨#OpenAI commits $1 billion to bolster cybersecurity for critical service providers via its “Daybreak for Frontline Defenders” initiative - targeting water utilities, grids, governments, and more. #Cybersecurity #AI