Meta, the social media giant known for its expansive digital empire, is currently at the center of a controversy involving claims of ‘open washing.’ This allegation follows the company’s sponsorship of a research paper published by the Linux Foundation that discusses the growth of open source AI solutions.
The study, released earlier this month, underscores the advantages of open source AI systems, particularly their cost-effectiveness for enterprises of all sizes, including small businesses. It highlights that organizations utilizing closed source AI models could expect to spend approximately three and a half times more than their open source counterparts.
However, critics argue that Meta’s promotion of its ‘Llama’ AI models as ‘open source’ does not adhere to the true definitions of open source. Amanda Brock, CEO of OpenUK, pointed out that Llama fails to meet the standards set by the Open Source Initiative because of licensing restrictions that limit its usability for certain purposes. She emphasized that true open source should be freely usable by anyone and for any purpose, which she claims Llama does not satisfy.
While Meta’s models have been marketed as open source, industry stakeholders, including the Open Source Initiative, have challenged this claim, citing restrictions on commercialization that conflict with open source principles. Earlier this year, the OSI explicitly criticized Meta for falsely promoting Llama as open source.
This ongoing dispute over licensing definitions extends beyond Meta. In March 2024, Databricks launched its own large language model, DBRX, which also failed to meet open source standards due to external use policies and licensing complications.
Despite these criticisms, Meta has taken steps towards open source awareness, though critics argue that the industry needs clearer standards to prevent ‘open washing’—a practice where companies market products as open source without fulfilling the necessary criteria.
Experts warn that open washing is not only an industry issue but a societal concern, especially as regulators in regions like the European Union begin to craft policies based on open source definitions for AI liability and standards.
In conclusion, while Meta’s contribution to open source AI is significant, the debate over licensing and transparency continues. It serves as a reminder that in the rapidly evolving tech landscape, clear standards are essential to foster genuine innovation and trust.
