Tarik ESSAADI's Highlights on 'Two models of AI oversight - and how things could go deeply wrong' | Glasp