**[OpenAI's New Roadmap Revealed: AI to Handle Most of Its Own R&D by 2028]** According to monitoring by Beating, OpenAI co-founder Sam Altman and Chief Scientist Jakub Pachocki co-authored an article unveiling the company's long-term roadmap, announcing the transition into the third development phase, centered on accessibility and safety. The roadmap emphasizes that the ultimate goal of artificial intelligence is to serve humanity and decentralize power, rather than replace human decision-making. The roadmap establishes three long-term objectives: building automated AI researchers, accelerating global economic growth, and enabling everyone to have their own personal AGI. To address the challenge of model alignment, OpenAI plans to achieve collaboration between AI systems and researchers to jointly complete most R&D work by March 2028. Sam Altman warned that fully automating everything not only lacks a sense of accomplishment but is also highly dangerous. The core role of humans remains to decide which endeavors are worth pursuing. At the operational level, OpenAI emphasized the need to prevent intelligent technologies from leading to excessive concentration of power. To counteract this, the roadmap reiterates the necessity of establishing a global coordination organization similar to the International Atomic Energy Agency. This organization would reduce catastrophic risks by sharing safety standards and, when necessary, collaborate to slow down the development of cutting-edge technologies. As the global economy is reshaped around AI, the new phase focuses on improving system usability and safety, transforming cutting-edge capabilities into universally accessible tools. [Original Link]