{"@context":"https://neupai.io/schema/v0.2","@type":"StructuredNewsArticle","identity":{"article_id":"tech42_20260429_ai-safety-discussion-expansion","canonical_url":"https://www.tech42.co.kr/ai-%ec%95%88%ec%a0%84%ec%84%b1-%eb%85%bc%ec%9d%98-%ec%9d%b4%ec%a0%9c-%eb%ac%bc%eb%a6%ac-%ec%84%b8%ea%b3%84%eb%a1%9c-%ed%99%95%ec%9e%a5%ec%97%90%ec%9e%84%ec%9d%b8%ed%85%94/?utm_source=rss&utm_medium=rss&utm_campaign=ai-%25ec%2595%2588%25ec%25a0%2584%25ec%2584%25b1-%25eb%2585%25bc%25ec%259d%2598-%25ec%259d%25b4%25ec%25a0%259c-%25eb%25ac%25bc%25eb%25a6%25ac-%25ec%2584%25b8%25ea%25b3%2584%25eb%25a1%259c-%25ed%2599%2595%25ec%259e%25a5%25ec%2597%2590%25ec%259e%2584%25ec%259d%25b8%25ed%2585%2594","ai_url":null,"publisher":{"name":"테크42","domain":"tech42.co.kr","type":"online"},"author":"황정호 기자","published_at":"2026-04-29T00:22:52.000Z","updated_at":null,"language":"en","article_type":"straight_news","originality":"self_produced"},"content":{"headline":"\"AI Safety Discussions Now Expanding to Physical World\"...AIM Intelligence Participates in Cooperation Discussions Between MSIT and Google DeepMind","summary":"The Ministry of Science and ICT and Google DeepMind signed an MOU for cooperation in AI joint research and talent development, with AIM Intelligence participating in discussions on AI safety assessment and standardization. It is noteworthy that AI safety discussions are expanding from text-based to agent and physical AI safety.","topics":["ai safety","government policy","international cooperation","technology standardization"],"geography":["KR"],"entities":[{"name":"Ministry of Science and ICT","canonical_id":"org:kr:ministry-of-science-and-ict","type":"organization","role_in_article":"primary_subject","metadata":{"ticker":null,"parent":null}},{"name":"Google DeepMind","canonical_id":"org:us:google-deepmind","type":"organization","role_in_article":"primary_subject","metadata":{"ticker":null,"parent":null}},{"name":"AIM Intelligence","canonical_id":"corp:kr:aim-intelligence","type":"company","role_in_article":"primary_subject","metadata":{"ticker":null,"parent":null}},{"name":"Park Ha-eon","canonical_id":"person:kr:park-ha-eon","type":"person","role_in_article":"quoted","metadata":{"ticker":null,"parent":null}},{"name":"Bae Kyung-hoon","canonical_id":"person:kr:bae-kyung-hoon","type":"person","role_in_article":"quoted","metadata":{"ticker":null,"parent":null}},{"name":"Demis Hassabis","canonical_id":"person:uk:demis-hassabis","type":"person","role_in_article":"quoted","metadata":{"ticker":null,"parent":null}},{"name":"Korea AI Safety Institute","canonical_id":"org:kr:korea-ai-safety-institute","type":"organization","role_in_article":"mentioned","metadata":{"ticker":null,"parent":null}}],"claims":[{"id":"c1","statement":"The Ministry of Science and ICT and Google DeepMind have embarked on cooperation for AI joint research in science and technology, talent development, and responsible AI utilization","as_of":"2026-04","as_of_explicit":false,"as_of_raw":"April 2026","source_type":"government_official","comparison":null,"type":"fact","figures":null,"expiry_hint":null,"insight":null},{"id":"c2","statement":"AIM Intelligence is an AI safety and security specialized company established in 2024","as_of":"2024","as_of_explicit":true,"as_of_raw":"2024","source_type":"company_disclosure","comparison":null,"type":"fact","figures":null,"expiry_hint":null,"insight":null},{"id":"c3","statement":"The joint research on text-to-video model jailbreaking techniques in which CTO Park participated was accepted to the ICLR 2026 main track","as_of":"2026","as_of_explicit":true,"as_of_raw":"ICLR 2026","source_type":"company_disclosure","comparison":null,"type":"fact","figures":null,"expiry_hint":null,"insight":null},{"id":"c4","statement":"The scope of AI safety discussions is rapidly shifting from text output safety to AI agent behavior and physical AI safety","as_of":"2026-04","as_of_explicit":false,"as_of_raw":"now","source_type":"company_plan","comparison":"previous_period","type":"opinion","figures":null,"expiry_hint":null,"insight":null},{"id":"c5","statement":"The standards for evaluation systems and safety measures themselves need to be newly established","as_of":"2026-04","as_of_explicit":false,"as_of_raw":"April 2026","source_type":"company_plan","comparison":null,"type":"opinion","figures":null,"expiry_hint":null,"insight":null}],"ai_emotional_context":{"valence":0,"arousal":0,"primary_emotions":[],"secondary_emotions":[],"emotional_triggers":[]}},"provenance":{"source_chain":["primary_reporting"],"original_source_url":null,"related_articles":[]},"temporal":{"freshness":"recent","next_update_expected":null},"access":{"license":"neupai_standard","attribution_required":true,"structured_data":"free","full_text_available":false,"full_text_access":null}}