Anthropic warns in S-1A filing that it is "very close" to achieving recursive self-improvement (RSI) capability, or "self-building AI". The AI lab stated in its updated S-1A filing on Monday that it may "imminently meet" its own internal definition of "dramatically accelerated progress attributable to AI," indicating it is on the cusp of creating a model capable of RSI. The company cautions that using AI to build more capable AI could pose a considerable safety risk, particularly if it sharply accelerates the pace of progress. However, Anthropic also noted that none of its current models have crossed this RSI threshold yet and it plans to keep humans in the R&D process for the foreseeable future.