OpenAI Warns Superintelligence Could Concentrate Power Without Decentralization

OpenAI published five guiding principles on April 26, warning that superintelligence could consolidate power among a small group of companies. The lab pledged to widely disseminate the technology to prevent that outcome. Sam Altman shared the framework on X. It replaces OpenAI’s 2018 AGI charter and lands as decentralized AI projects compete for the same narrative. "Power in the future can either be held by a small handful of companies using and controlling superintelligence, or it can be held in a decentralized way by people. We believe the latter is much better, and our goal is to put truly general AI in the hands of as many people as… https://t.co/hOx6M4jH3r— OpenAI Newsroom (@OpenAINewsroom) April 27, 2026 OpenAI Reframes Superintelligence Around Five Principles The five principles are democratization, empowerment, universal prosperity, resilience, and adaptability. The first commits OpenAI to resisting any concentration of AI control, including within the company itself. It also routes key decisions through democratic processes rather than internal lab choices. Altman framed it as the lab’s first major principles update since 2018. Empowerment promises broad public access to general AI and the tokens markets that have grown around it. The remaining three pillars cover economic transition risks, coordination on safety, and a willingness to revise positions. The 2026 charter mentions AGI only twice, signaling a shift toward a wider commitment to AI infrastructure. Our Principles:Democratization, Empowerment, Universal Prosperity, Resilience, and Adaptabilityhttps://t.co/g58WGndARw— Sam Altman (@sama) April 26, 2026 Decentralized AI Rivals Push Back The warning lands as crypto-native AI networks expand. Bittensor (TAO) ran the largest-ever decentralized large-language-model training on its Templar subnet in early April. Grayscale has filed for a TAO-focused ETF, drawing fresh institutional capital to the network. Critics argue that OpenAI is raising concerns ab...