Meta is under fresh criticism for what experts are calling “open washing” after sponsoring a Linux Foundation whitepaper promoting the benefits of open-source AI.
The paper argues that open models can significantly cut enterprise costs, claiming proprietary AI tools cost more than three times as much. However, Meta’s role in the report has sparked controversy, particularly over its Llama AI models, which the company markets as open source—despite licensing restrictions.
Amanda Brock, CEO of OpenUK, said Llama’s license violates open-source standards set by the Open Source Initiative (OSI), citing restrictions on commercial use. “Open source should allow unrestricted use, and Llama does not meet that bar,” Brock said.
Though Meta has positioned Llama as open source, OSI and other experts argue the licensing terms undermine core principles of openness and accessibility.
Brock acknowledged Meta’s role in fostering conversation around open source but warned that mislabeling models could mislead lawmakers and regulators shaping AI policy.
Meta is not alone. Databricks faced similar scrutiny in 2024 over its DBRX model, also criticized for falling short of OSI standards. As AI adoption accelerates, the distinction between truly open and merely accessible remains a flashpoint for debate.
