The third report of the One Hundred Year Study on Artificial Intelligence (AI100) project, conducted under the auspices of Stanford HAI, assesses that artificial intelligence is developing faster than anyone expected and has outpaced existing governance frameworks. The study panel for 2026 was led by Michael Wooldridge of the University of Oxford, and experts in computer science, philosophy and sociology took part in its preparation.
According to the report, the main driver of progress so far, scaling models through ever larger amounts of data and computing power, is reaching its limits. High quality data is becoming scarcer, and the costs of training top models are becoming too high. The next phase, the researchers say, will likely be marked by smarter algorithms, more efficient models and systems capable of deeper reasoning.
The report warns that disruptions are spreading to areas long considered resistant to automation, such as creative industries, software development and mathematics. AI is also accelerating scientific research, shortening timelines from years to months or weeks, but it requires new methods for verifying machine generated discoveries.
Special attention is devoted to the investment boom. Capital is flowing into AI infrastructure at a pace that surpasses earlier technology cycles, data center construction is straining energy systems, and chipmaker NVIDIA briefly became the most valuable company in history. However, the report notes that the economic effects of AI are not yet visible in aggregate data or in return on investment, which raises concerns about a possible financial bubble.
The development of frontier AI is concentrated among a small number of actors in the US and China, while military applications have moved from theory into practice, including decisions on lethal targets without human oversight. Generated content, according to the report, has undermined trust in photographic evidence, and international governance of AI remains fragmented.
What this means for businesses
For companies, the report signals that competitive advantage increasingly depends less on model size and more on efficiency, data quality and reasoning capability. Rising infrastructure costs and unclear returns on AI investment justify caution in budget planning and in assessing the risk of a stock market correction. Firms that use generated content or voice technologies should strengthen identity checks and internal controls, because misuse is becoming a real business threat. Fragmented regulation means that compliance will depend on the markets in which a company operates.
This summary was written with AI based on the cited source and checked by automated rules. For exact details, read the original article.