Facebook founds FAIR with Yann LeCun as director

Background

By the end of 2013 Facebook wanted its own research organisation for artificial intelligence, the field that builds systems able to learn from data and perform tasks that once needed human judgement. Mark Zuckerberg took the lead in assembling the team.^[1]

What happened

Facebook launched Facebook AI Research (FAIR) in late 2013. Yann LeCun, a professor at New York University, became FAIR's first director on 9 December 2013 in New York City. He kept his professorship while leading the lab.^[2]

Zuckerberg personally went to the NeurIPS conference to recruit researchers. LeCun partnered with him to assemble the founding team of deep-learning researchers.^[1]

Key details

Meta later described FAIR's purpose as advancing the state of the art in AI through open research.^[1]

FAIR work that Meta highlights includes Faster R-CNN (2015), Mask R-CNN (2017), PyTorch (2017), Segment Anything (2023) and the Llama models (2023).^[1][3]

LeCun was succeeded as FAIR director by Jérôme Pesenti in 2018. After Facebook became Meta in 2021, FAIR was renamed Fundamental AI Research.^[3]

A research lab of this kind publishes its results, so other researchers can check, reuse and extend them. That is what Meta means by open research: the work is shared outside the company rather than kept inside it.^[1]

The founding team was built around deep-learning researchers. Deep learning uses neural networks with many layers to learn patterns from large amounts of data, and that is the approach behind the lab's later framework and model work.^[1]

Impact

In 2025 FAIR was folded into Meta Superintelligence Labs. LeCun announced on 19 November 2025 that he was leaving Meta after about ten years.^[4][2]

The lab's history connects the early open research culture to later products and models. Its founding team and its early projects are part of the background to the frameworks and model families described in other entries in this archive.