Search results for AI Safety Research Institute
AI & Enterprise
Tech Insight: Current way of building AI models cannot solve AI safety problems
Yoshua Bengio (요슈아 벤지오), a Turing Award winner and one of the world’s most-cited scientists, shared views on why AI agent malfunctions are increasing and how to respond. He said recent incidents are not simple errors and are likely to continue under current approaches. Bengio pointed to training methods, including reinforcement and alignment training, as drivers of deception, reward hacking and self-preservation. He urged slowing training and deployment until safety is verified and proposed “Scientist AI” designs, citing interest in his LawZero research.
Games & Commerce
Kakao steps up AI ethics governance, launches external advisory panel
Kakao said on Wednesday it has launched an external advisory panel under its AI and technology ethics subcommittee to strengthen its AI ethics governance. The move aims to pre-empt potential social risks from the spread of agentic AI and to establish internal policies in line with fast-changing domestic and international environments, including Korea’s AI basic law and global AI safety regulation. The panel consists of seven external experts spanning academia, law and an international organisation.
AI & Enterprise
MSIT holds cybersecurity workshop based on OpenAI TAC
South Korea’s Ministry of Science and ICT said it held an AI security hands-on workshop with OpenAI on May 18, joined by several government bodies and related institutions. OpenAI’s Sasha Baker explained the Trusted Access for Cyber programme and demonstrated cybersecurity functions of the latest AI models. The ministry asked OpenAI to share information and cooperate on responding to AI security threats, and both sides agreed to continue working-level discussions.
