Trust and believe – AI models trained to see ‘legal’ doc as super legit Researchers at security firm Pangea have discovered yet another way to trivially trick large language models (LLMs) into ignoring their guardrails. Stick your adversarial instructions somewhere in a legal document to give them an air of unearned legitimacy – a trick familiar to lawyers the world over.… READ MORE HERE…

The admin controls were left wide open on Pudu’s robots A researcher caught the world’s leading supplier of commercial service robots using shoddy admin security that let attackers redirect the delivery machines to anywhere and make them follow any command.… READ MORE HERE…

Look who’s visiting the watering hole these days Amazon today said it disrupted an intel-gathering attempt by Russia’s APT29 to trick Microsoft users into unwittingly granting the Kremlin-backed cyberspies access to their accounts and data.… READ MORE HERE…

Up to 29,000 organizations and potentially 370,000 security and IT pros affected Australian development house Click Studios has warned users of its Passwordstate enterprise password management platform to update immediately if not sooner, following the discovery of an authentication bypass vulnerability that opens the doors to an emergency administration account with nothing more than a “carefully crafted URL.”… READ MORE HERE…