{"type":"video","version":"1.0","html":"<iframe src=\"https://www.loom.com/embed/96a3965921fd408bae410146e5b696ec\" frameborder=\"0\" width=\"1280\" height=\"960\" webkitallowfullscreen mozallowfullscreen allowfullscreen></iframe>","height":960,"width":1280,"provider_name":"Loom","provider_url":"https://www.loom.com","thumbnail_height":960,"thumbnail_width":1280,"thumbnail_url":"https://cdn.loom.com/sessions/thumbnails/96a3965921fd408bae410146e5b696ec-2d1daa97933fbee0.gif","duration":372.633333,"title":"AI Compliance Risk: Nested Tech Taxonomy","description":"This Loom explains a regulatory and governance taxonomy for AI by mapping how artificial intelligence relates to machine learning, deep learning, and generative AI. It emphasizes these technologies are nested subsets rather than competing products, and that placing a system in the correct layer determines its compliance risk vector. The video notes AI can be rule based and does not need to learn, while machine learning relies on training data and is vulnerable to bias, deep learning adds opacity from multi layer neural networks, and generative AI creates new content and raises risks like hallucinations, copyright and IP exposure, and deepfakes. It also traces the timeline from AI in the 1950s, machine learning breakthroughs in the 1980s and 1990s, deep learning in 2012, Transformers in 2017, to ChatGPT in 2022, concluding that terminology is necessary for applying frameworks like NIST and ISO."}