AI "Godfather" Geoffrey Hinton and 21 other top AI scholars have jointly published the first paper on Recursive Self-Improvement (RSI), warning that AI's full takeover of R&D could trigger an "intelligence explosion." The paper states that once a critical point is crossed, AI advancements that take a year today could, in extreme cases, take only about five weeks in the future. Data from Anthropic shows that the proportion of AI-driven AI R&D work has increased from 1% in March 2026 to 26% in August. The paper also warns that the growth in capabilities brought about by AI R&D automation could far outpace society's ability to adapt, threatening to erode human control over AI systems and potentially disrupting the balance of power between nations, corporations, and government agencies. It calls on policymakers to immediately establish regulatory frameworks.