Microsoft's AI Frontiers Is Gone, Founding Researcher Says as He Departs
In a November 6 post, John Langford said AI Frontiers had come to an end and that he had left Microsoft. The account follows the company's September announcement that the group was simply moving and its research would continue.
John Langford, who helped found Microsoft’s AI research group AI Frontiers, says the organization is finished and that he has left the company. Langford laid out what happened after the reorganization in a November 6 post. By his account, AI Frontiers had roughly 50 people. The change Microsoft announced publicly was that AI Frontiers would move from Microsoft Research to Microsoft AI.
What happened after the move
Microsoft announced the reassignment of AI Frontiers on an official page on September 14. At the time, the company said the group would continue its work on efficient language models and on AI that handles tasks on its own, known as agents, under the new organization. The message was that the group was moving to Microsoft AI while continuing its research.
The situation Langford describes does not match that. A researcher who helped found the group has now announced both its end and his own departure.
The two accounts also diverge on research direction. Langford argues that Microsoft AI’s research priorities did not line up with those of the existing team. In its September announcement, Microsoft said the new organization was a good fit for continuing the work.
What happens to the existing research
AI Frontiers was a group focused on models and agents. Microsoft points to AutoGen, Fara, and Magentic-One as research output from the team.
For anyone evaluating or already using those projects, the important question is what form the research takes from here. Who will be responsible for each project and what results they release next matter more than the group’s name or organizational home. Developers and companies in Korea weighing adoption will likewise want to confirm how the existing projects will be maintained and what follow-up research is planned.
Two things are worth watching. How Microsoft describes the current status of AI Frontiers, and what follow-up plans it offers for existing work such as AutoGen.