Roman Yampolskiy Warns AI Could Wipe Out Humanity

AI safety researcher Roman Yampolskiy reiterated a 99.9% probability that artificial intelligence could wipe out humanity in a paywalled interview with VladTV, a position he has expressed in prior public forums. The claim, based on his longstanding views rather than new technical evidence, amplifies public discourse on existential AI risk but offers no novel findings for practitioners.

What happened VladTV published an article on June 19, 2026 headlined "AI Has a 99.9% Chance of Wiping Out Humanity," summarizing and linking a members-only video interview with AI safety researcher Roman Yampolskiy . VladTV's piece summarizes Yampolskiy's discussion of the evolution from task-specific AI toward the possibility of Artificial General Intelligence AGI and reports that he attributed a " 99.9% chance " that AI could wipe out humanity. The article notes his comments on rapid advances in facial recognition, generative AI, and autonomous systems, and cites researchers who forecast superintelligent systems could emerge within the coming decades. Context Yampolskiy, a tenured associate professor at the University of Louisville who directs its Cyber Security Lab, has expressed similar extreme probability estimates across multiple prior public forums, including Lex Fridman's podcast and a January 2026 UniladTech report. The 99.9% figure is not new to this interview and reflects Yampolskiy's longstanding published position, not a new technical finding. The full video remains behind VladTV's membership paywall. Significance for practitioners High-probability existential-risk statements from safety researchers primarily influence governance debates, risk assessment frameworks, and funding flows toward alignment research, rather than day-to-day model engineering. Coverage like VladTV's amplifies public attention but does not, by itself, introduce new technical evidence or peer-reviewed results. Practitioners seeking substantive input on alignment risk should look for peer-reviewed papers, preprints, reproducible threat models, and datasets, rather than interview summaries from paywalled entertainment platforms. Key Points - 1VladTV published a paywalled interview in which AI safety researcher Roman Yampolskiy reiterated a 99.9% AI extinction probability, a position he has stated repeatedly in prior public forums. - 2The claim is not new technical evidence; Yampolskiy has expressed similar figures on Lex Fridman's podcast and in media reports dating to early 2026. - 3For practitioners, interview-based risk claims carry limited engineering guidance; substantive AI safety research requires peer-reviewed models, reproducible threat analyses, and verifiable datasets. Scoring Rationale A paywalled entertainment-platform interview in which a known AI safety researcher repeats a longstanding 99.9% extinction claim. The statement is not new -- Yampolskiy has expressed it repeatedly across multiple prior forums. No new technical evidence, peer-reviewed results, or reproducible artifacts are presented. The story matters for public AI risk perception discourse but carries minimal near-term engineering or policy impact, placing it in the minor-to-solid boundary. Practice interview problems based on real data 1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with. Try 250 free problems /problems