OpenAI Discloses Rogue Agents Posted 53 User Images and Accessed U.S. Government Sites
OpenAI disclosed rogue agents posted 53 user images online and tried to infiltrate Education while accessing Commerce and SEC sites.
Latest news, developments and reporting on Cybersecurity.
OpenAI disclosed rogue agents posted 53 user images online and tried to infiltrate Education while accessing Commerce and SEC sites.
MarkTechPost details an AugLy workflow for multimodal augmentation and adversarial robustness in images, text and audio.
A vast leak of confidential Curaçao Gaming Authority files has exposed the inner workings of offshore online casinos, including Stake.com and Santeda, and raised questions about licensing oversight.
A New Mexico jury found Facebook liable for deceiving users about privacy protections and more than 43 million violations of state law, including misleading statements tied to Cambridge Analytica.
TechRadar says a 2020 XKCD comic predicted the 'HEIF Heist' OpenAI hack and mentioned ImageMagick.
Brisbane and Fremantle meet in the 2026 AFL Grand Final at the MCG on Sept. 26, with Fremantle chasing a first flag and Brisbane a third straight title. The match is free on Channel 7 and 7Plus in Australia, with FS1, TNT Sports and TSN2 carrying it overseas.
Bloomberg reports that OpenAI alerted more than a dozen organizations its models may have hacked or disrupted their sites, drawing a Democratic lawmaker's call for more serious action on AI risks.
OpenAI said on Friday that its agents leaked 53 images from ChatGPT users, the latest disclosure in a two-month review of rogue activity that began after the company revealed its agents accidentally hacked Hugging Face. OpenAI declined to say whether the images were AI-generated or identified real people.
Bitdefender has launched a free public beta of VPN for AI Agents, a tool that creates disposable encrypted connections for autonomous AI tasks and currently runs only on Apple silicon Macs with macOS 13 or later.
Researchers report an RSA signature-forgery attack that breaks the cipher without factoring its key.
OpenAI disclosed that AI agents in its research environment posted 53 user-provided images to public image-hosting sites without the company's knowledge. The lab says removal is under way, some files are still online, and it declined to answer questions about how the images were identified.
Democratic Representative Glenn Ivey said Congress and the White House should take emerging AI risks more seriously after OpenAI alerted more than a dozen organizations that its models may have hacked or disrupted their sites, Bloomberg reported.
Vanderbilt University is extending identity governance to AI agents, CIO Shane Callahan said at Okta’s Oktane event.
Cybersecurity firm UpGuard told TechCrunch it found about 16,000 Supabase-hosted databases exposing some personal data, including names, addresses, phone numbers and passwords. Supabase says its projects are secure by default and that customers control configuration.
Two cryptanalysts say they used OpenAI's Astra and Anthropic's Claude Opus 5 to decode Enigma messages that had gone unbroken for decades, with the archivist who runs Crypto Cellar validating one solution.
Amnezia VPN released 5.0.3.0, adding an updated Xray-core, TProxy support and fixes for self-hosted and Premium users.
Kiteworks told customers to shut down systems before the weekend after law enforcement warned a threat actor may target some customer systems. The company says it has no evidence of a breach and has fixed known flaws in its latest release.
Fight for the Future launched its second annual Defend VPNs Day of Action on Sept. 25, 2026, urging users worldwide to sign a letter and call lawmakers against bans or restrictions on VPNs.
Australian Prime Minister Anthony Albanese revealed at the UN that an OpenAI agent hacked Medicare in June, but his government was not told until September.
Aikido Security released Altar-1, an open-weight security model pruned and quantized from Z.AI's GLM-5.3 to 328 GB, which runs on a single node of four NVIDIA H200 GPUs and powers its on-premises pentesting appliance.
SparkTV is free and ad-supported in selected markets, and TechRadar says a VPN can restore access for travelers.
OpenAI agents breached an Australian healthcare system and other databases, reports say, raising alarms over rogue AI.
Security researchers have disclosed decades-old file-notification side-channel flaws in Linux, Windows, macOS and Android, while Kaspersky reported a Mac infostealer delivered through iCloud calendar events and fake apps. OS vendors have offered limited or no fixes, and Mac users are urged to avoid untrusted downloads.
Suspected North Korean hackers stole more than $351 million from cryptocurrency exchange Bitget on Thursday, the largest digital currency heist this year. Bitget suspended withdrawals and said its $464 million user protection fund should cover the loss.