From AGI to ASI

Over the last decade, building human-level artificial general intelligence has moved from far-fetched speculation to being a concrete next-decade target for many of the largest AI organisations. Achieving this goal would have profound and far-reaching impacts on human society, which raises many complex questions for the decade ahead. This report investigates how AI itself might continue to develop in a post-AGI world along the continuum of machine intelligence. The endpoint of this continuum, Universal AI, is theoretically well understood, which provides some formal grounding for the main focus of this report: the transition from human-level AGI to artificial general superintelligence, which, intuitively, can be understood as a system that is more intelligent and cognitively capable than large organisations of humans. After characterizing ASI, the report discusses four potential pathways from AGI to ASI: scaling AGI, AI paradigm shifts, recursive improvement, and ASI emerging from large-scale multiagent collectives. The report then discusses possible frictions and bottlenecks along these pathways. Determining whether the impact of these frictions will be negligible or substantial raises a number of concrete open research questions.

Introduction. The main goal of this report is to take a close look at AI progress beyond human-level AGI (independent of when humanity hits this milestone), and to map out a landscape around potential technological pathways for continued AI progress, as well as possible frictions that would slow down (or halt) progress along these pathways. Determining the impact of these frictions leads to concrete open research questions. The future is unpredictable. This is true for predicting the pace of technological progress and for predicting how new technologies will affect society. Progress in AI has been very rapid over the past decade, lending urgency to the question of how this progress will continue and what impact it will have on human society. The field is witnessing historically unprecedented amounts of compute, researchers, funding, and large-scale coordinated efforts. Extrapolating scaling and growth trends from the past decade leads to forecasts for the next decade that sound like science-fiction (Aschenbrenner, 2024; Kokotajlo et al., 2025; MacAskill and Moorhouse, 2025).

Discussion / Conclusion. As stated at the beginning of this report: “the future is unpredictable”, but we can be better prepared by reducing uncertainty through more landscape-mapping work like ours, having a large range of concrete speculative scenarios, and ramping up research efforts to study advanced AI systems, their properties, and potential impacts. One big lever is to ramp up efforts for developing robust and more reliable AI benchmarking and forecasting methods that continue to work in a post-AGI future. Instead of focusing on one technological trajectory and timeline, being prepared for a post-AGI world requires considering a diverse set of forecasts and scenarios, paired with continual benchmarking and monitoring to update the set of forecasts and scenarios and their relative plausibility.