Yann LeCun is a French-American computer scientist, Chief AI Scientist and Vice President at Meta, and Silver Professor at New York University (Courant Institute of Mathematical Sciences and Center for Data Science). He received the 2018 Turing Award jointly with Yoshua Bengio and Geoffrey Hinton for foundational work on deep learning. He is among the more prominent scientific skeptics of AI existential-risk arguments and an advocate of open-weight frontier AI.
Type: Individual (academic, industry researcher) Affiliations: Meta (Chief AI Scientist, VP); New York University (Silver Professor, Courant Institute and Center for Data Science) Notable recognition: 2018 Turing Award (with Yoshua Bengio and Geoffrey Hinton) for foundational deep learning work
Background
LeCun's work in the 1980s and 1990s on convolutional neural networks (CNNs) laid groundwork for modern computer vision; his LeNet architecture is a direct ancestor of contemporary image-recognition systems. He was a researcher at AT&T Bell Labs before joining Meta (then Facebook), where he was founding director of the FAIR (Facebook AI Research) lab in 2013. He led FAIR from 2013 to 2018, a period during which the lab's research-culture choices produced PyTorch and, later, the Llama series. He is now Chief AI Scientist at Meta, where he shapes the company's open-weight Llama strategy and longer-horizon research on non-autoregressive architectures, and a Silver Professor at NYU.
Contributions and works
LeCun pioneered convolutional neural networks in the 1980s and 1990s. The CNN work was set out in "Gradient-Based Learning Applied to Document Recognition" (LeCun, Bottou, Bengio, Haffner, 1998), a foundational CNN paper. He co-authored the survey "Deep Learning" (LeCun, Bengio, Hinton) in Nature in 2015.
At Meta he developed the Joint Embedding Predictive Architecture (JEPA) research agenda, presented as an alternative to pure autoregressive LLM scaling and set out in "A Path Towards Autonomous Machine Intelligence" (2022). He also publishes commentary on X/Twitter and is a widely-read contrarian voice in AI-risk discourse.
AI policy positions
LeCun's positions diverge from those of his Turing co-laureates Bengio and Hinton, particularly on existential risk.
On existential risk, he argues repeatedly that current-style LLMs are nowhere near AGI, that extinction scenarios are implausible, and that safety concerns reflect hype or corporate interest in regulatory capture. He has publicly characterised the existential-risk camp as "doomers," and did not sign either the FLI Pause Letter or the CAIS Statement on AI Risk.
On open source, he defends Meta's release of Llama weights as essential for research, safety auditing, and preventing closed-lab monopolies, a position aligned with the open-source AI policy camp.
On scaling, he argues that autoregressive LLMs cannot reason or plan adequately and that new architectures, including world models and JEPA, are needed to reach AGI.
On regulation, he opposes licensing regimes for frontier models and has publicly criticized proposals such as California SB 53-style frameworks, and especially the earlier SB 1047, as counterproductive.
Relationships
- contradicts: Yoshua Bengio, Geoffrey Hinton — Turing co-laureates; opposite views on existential risk
- contradicts: Statement on AI Risk (CAIS), FLI — Pause Giant AI Experiments: An Open Letter
- supports: Open-Source AI / Open-Weight Models
- related: AI as Normal Technology — LeCun's framing aligns with the normal-technology camp