Glossary
What is Mark Zuckerberg's role in AI?
In short
Mark Zuckerberg is the founder and head of an American technology company that releases the open Llama language models.
People · Mark Zuckerberg
MMOOZZGG.com editorsChecked: 1 min read
For a child
Imagine that at break time you built a tall tower of blocks and let all the other kids take its plan to build the same one at home. Mark Zuckerberg did something similar with smart programs that understand words. His team builds such programs and shares them with everyone, so other people can use them and change them in their own way.
In plain words
Zuckerberg founded Meta and runs it. In AI, Meta is known for its Llama models. The first version came out in February 2023 in four sizes, for researchers, and in July 2024 Zuckerberg published the letter “Open Source AI Is the Path Forward” and released Llama 3.1. His argument: open models will become a common base, as the Linux operating system once did. In June 2025 he put all the company’s AI teams into one unit. For example, a developer downloads the Llama weights and runs the model on their own computer without sending texts to someone else’s cloud.
For business
For a Russian company, the main interest of Zuckerberg’s work is Meta’s open Llama models. You can download their weights and run them on your own server, so customer data does not go to an outside vendor. This makes it easier to comply with 152-FZ, Russia’s personal data law, and removes dependence on someone else’s API. The costs move to your side: a server with GPUs, setup, updates and support. The license of each version sets the terms of use, so a lawyer reads it before rollout and also assesses the developer company’s legal status in Russia. Test the model’s quality separately in Russian and on your tasks. Which servers fit such models and what they cost to run is covered on the page about AI servers.
For a specialist
On 24 February 2023 Meta presented the first LLaMA (Large Language Model Meta AI): models with 7, 13, 33 and 65 billion parameters under a non-commercial research license; the 33B and 65B versions were trained on 1.4 trillion tokens. On 23 July 2024 Zuckerberg, as founder and CEO, published a letter about open AI and released Llama 3.1 405B, which he called the first frontier-level open model, together with updated 70B and 8B versions. In the letter he compares open models with Linux and writes that they give developers control over data and costs. On 30 June 2025 he announced Meta Superintelligence Labs: according to an internal memo, all the company’s AI teams move into the new group. It is led by Alexandr Wang, former head of Scale AI, as chief AI officer, and Nat Friedman, former head of GitHub, is in charge of products and applied research.
Example
A bank wants an assistant to answer staff questions on internal rules, but the texts must not go to an outside cloud. The team runs an open model on its own server, fine-tunes it on the rules and checks the answers on real staff questions.
Related terms
Questions and answers
Sources
Not sure whether your company needs this?
AI solutions map