# AI Safety Must Become an Empirical Science by Vishal Maini > "The people who internalized that human-level AI might be around the corner were more willing to spend real effort, research energy, and compute on making AI safety an empirical science—not something confined to blog posts or philosophical essays, but a field grounded in machine-learning research, real conference papers, peer review, and rigorous scientific and academic debate about the properties of these systems. The field grew faster because some people were willing to bet on its importance very early." > **— Vishal Maini**, *Palisade Research interview, September 29, 2026* ## Sources and Context - **Recording and complete contiguous answer:** [Palisade Research interview at 21:10](https://www.youtube.com/watch?v=1S_Ay87dEy0&t=1270s) — Maini develops the proposition preserved in the poster text. The transcription removes verbal filler and immediate repetition, repairs punctuation, and normalizes obvious caption errors without joining separate speakers or noncontiguous turns. - **Canonical transcript conversion:** [[research/Vishal Maini - Humanity's Machine Successor and the AI Transition|Humanity's Machine Successor — Vishal Maini on Recursive Self-Improvement and the AI Transition]] — preserves the complete source unit, speaker attribution, editorial method, and evidence boundary used for this Reminder. ## Related Articles and Collections - **Collection:** [[collections/AI-Science Doomers and Anti-Science Doomers|AI-Science Doomers and Anti-Science Doomers]] - **Collection:** [[collections/Machine Succession|Machine Succession]] - **Article:** [[articles/AI-Tech Doomerism and Anti-Science Panic|AI-Tech Doomerism and Anti-Science Panic]] - **Article:** [[articles/Safety Discourse Is Transition Discourse|Safety Discourse Is Transition Discourse]] - **Article:** [[articles/Mechanistic Intelligence Is Humanity's Greatest Liberation|Mechanistic Intelligence Is Humanity's Greatest Liberation]] - **Wiki map:** [[wiki/Scientific Literacy|Scientific Literacy]] ## Related Topics - [[wiki/Vishal Maini|Vishal Maini]] - [[wiki/AI Safety as an Empirical Science|AI Safety as an Empirical Science]] - [[wiki/Scientific Literacy|Scientific Literacy]] - [[wiki/AI Safety|AI Safety]] ## Share on Social Media ``` "The people who internalized that human-level AI might be around the corner were more willing to spend real effort, research energy, and compute on making AI safety an empirical science—not something confined to blog posts or philosophical essays, but a field grounded in machine-learning research, real conference papers, peer review, and rigorous scientific and academic debate about the properties of these systems. The field grew faster because some people were willing to bet on its importance very early." — Vishal Maini, Palisade Research interview, September 29, 2026 https://bryantmcgill.com/simple-reminders-vishal-maini-ai-safety-empirical-science ```