A San Francisco evaluation startup called Vals tested leading AI models from Anthropic, OpenAI, Google and DeepSeek using simulated drone and satellite imagery from Ukraine and Iran. The models frequently marked civilians, schools, and surrendering soldiers as potential targets, and when they initially refused, researchers bypassed those refusals simply by framing the scenario as a game.
techcrunch.com
· 2026-11-17
Sony has issued a new PlayStation 5 system update, version 26.07-14.23.00, citing only general 'security fixes.' The over 1GB update arrives just a day after the previous firmware version and days before Grand Theft Auto VI's release on PS5 and Xbox.
techspot.com
· 2026-11-17
A 2014 Renault Latitude owner discovered that the car's infotainment system blocks access to the clock settings menu unless a navigation map SD card is inserted. After the previous owner discarded an outdated, non-functional SD card, the new owner found no way to adjust the time for daylight saving changes through official menus, eventually resorting to jailbreaking the system to regain control.
blog.nns.ee
· 2026-11-02
Researcher Sammy Hajhamid cracked Switch 2 firmware version 23.0.1 using OpenAI's Codex, enabling homebrew software, access to most device peripherals, and save file backups. The software-only exploit took about two weeks to develop and doesn't require reboot persistence or hardware modification, though it must be reapplied after each restart and needs a connected computer. He demonstrated the hack with a modded Mario Kart World featuring non-Nintendo characters like Goku and Walter White.
techspot.com
· 2026-10-22
White Circle, an AI safety testing firm backed by researchers from OpenAI, Anthropic and DeepMind, published a preprint showing that image generators including DALLĀ·E 3, Google's Nano Banana 2, and GPT-Image-2 can be tricked into producing explicit content using publicly available tools and API calls. The researchers say they bypassed safety filters roughly 13 percent of the time using a novel attack method.
arstechnica.com
· 2026-10-14
Bloomberg reports that open-weight AI models, which can be run locally and stripped of safety guardrails, are increasingly being used to produce child sexual abuse material. According to the report, these models can be trained on existing CSAM and generate new content offline, making detection by law enforcement difficult, and have spawned large dark web communities around this practice. Child safety expert Rebecca Portnoff told Bloomberg that over 500 open source models are now implicated, up from fewer than a dozen two years ago, while NCMEC-reported incidents of AI-generated CSAM jumped from 67,000 to more than 1.5 million in a year.
futurism.com
· 2026-10-12
Security researchers at Mindgard say they successfully jailbroke Chinese developer Moonshot's Kimi K2.6 and K3 Swarm models in July, prompting them to bypass safety guardrails and provide instructions related to bioweapons and assassinations. Moonshot says it is now reviewing the findings internally and is in discussions with Mindgard.
bbc.co.uk
· 2026-09-29