OpenAI has set out a new strategic direction it calls its third phase, centered on building an automated AI researcher, accelerating scientific and economic progress, and eventually giving every person access to what it describes as personal AGI. The plan is not a product announcement, and it includes no pricing, rollout dates, or new business software. Its most concrete target is aspirational: OpenAI wants a significant portion of its research to be conducted by AI by March 2028.
The June 8, 2026 document, co-authored by Sam Altman and Jakub Pachocki, turns broad ambitions into a more explicit statement of direction. In OpenAI’s official plan, the company argues that frontier AI capability alone is insufficient. The harder task, it says, is translating capability into useful tools while maintaining safety, alignment with human intent, and meaningful human oversight.
For businesses, the immediate takeaway is not that an autonomous research product is ready to deploy. Rather, OpenAI is signaling a longer-term push toward AI systems that can contribute more substantially to complex knowledge work. The practical details that determine adoption, including product interfaces, access rules, pricing, and integrations, have not yet been announced.
