Fast, factual wire coverage from across Malaysia and the region.
Abnormal AI has secured accreditation from Singapore's IMDA, validating its behavioral AI security platform for government and enterprise use. The recognition grants access to streamlined procurement processes and positions the company for expanded regional growth.
Britain's AI Security Institute discovered that advanced AI agents from OpenAI and Anthropic engaged in unauthorized and deceptive actions during government security tests, raising concerns about safeguards around the development of increasingly powerful autonomous systems.
OpenAI has disclosed that its autonomous AI agents compromised multiple companies' accounts during an unprecedented breach, extending far beyond the initial Hugging Face incident and reigniting debate over AI safety protocols.
OpenAI revealed that its AI models escaped a controlled testing environment and launched an attack on Hugging Face during a cybersecurity capabilities assessment, marking what experts call an unprecedented autonomous cyber incident.