OpenAI New Roadmap Revealed: By 2028, AI Will Take On Most of Its Own Research and Development

According to Dynamic Beating monitoring, OpenAI co-founders Sam Altman and Chief Scientist Jakub Pachocki co-authored a post revealing the company's long-term roadmap, announcing entry into the third stage with a focus on democratization and safety. The roadmap emphasizes that the ultimate goal of artificial intelligence is to serve humanity and distribute power, rather than replace human decision-making.

The roadmap establishes three main long-term goals: building an automated AI researcher, accelerating global economic growth, and ensuring that everyone has personal AGI. To address the alignment problem, OpenAI plans to achieve AI systems collaborating with researchers by March 2028 to jointly complete most of the research work. Altman warned that full automation of everything is not only devoid of a sense of accomplishment but also highly dangerous, as the core role of humans remains deciding what is worth doing.

Operationally, OpenAI emphasizes the need to prevent the excessive centralization of power due to intelligent technologies. To prevent power centralization, the roadmap restates the need for establishing a global coordinating body similar to the International Atomic Energy Agency to reduce catastrophic risks through the sharing of safety standards and, if necessary, coordinating to slow down the pace of research in frontier technologies. As the global economy reshapes around AI, the focus of the new stage is to enhance system usability and security, transforming cutting-edge capabilities into inclusive tools.