Dr Waku

Dr Waku is a pseudonymous AI safety educator and content creator based in London. The name serves as both a personal brand and an informal research identity. Dr Waku holds a PhD in computer security from an Ivy League institution and works professionally as an AI research scientist, bringing an adversarial security mindset to questions of AI safety. Dr Waku became interested in AI after a chronic health condition limited their typing ability, prompting the development of voice recognition and voice coding systems. This personal experience with assistive AI deepened their engagement with the field and eventually led to public education work. The primary output is a YouTube channel with approximately 30,500 subscribers as of early 2026, focused on educational content about AI and its societal implications. Topics include why AI alignment is hard, outer and inner misalignment, reward hacking, deceptive alignment, recursive self-improvement risks, and AI security vulnerabilities. Dr Waku also maintains a Substack newsletter (launched October 2024) covering similar themes, and has appeared on podcasts including the ControlAI Podcast. Dr Waku has been involved with the AI safety community through Manifund (donating to projects including AI Safety Camp and the AI Governance Exchange), and participated in AI Safety Camp 11 with a project focused on creating YouTube Shorts content about AI loss-of-control risk. Content also appears via blog.biocomm.ai and other AI safety media outlets. The operation is currently run part-time alongside other employment, with a small Patreon providing minimal supplemental income. There is no formal organizational structure, staff, or significant institutional funding.