AI in practice

Meta continues to advocate for open source AI

Matthias Bastian
Mit dem halb veröffentlichten, halb geleakten LLaMA-Modell hat Meta eine Open-Source-Bewegung für Sprachmodelle in Gang gesetzt. Auch in Zukunft will Meta offen agieren.

Midjourney prompted by THE DECODER

In a submission to the NTIA, Meta argues that open-source AI leads to better and safer products, faster innovation, and a larger market. Open models would democratize access to AI and create opportunities for everyone, not just technology giants. Closed models often only appear secure because vulnerabilities can go undetected for a long time. If a closed model is hacked, only the model developer has the information needed to mitigate the damage, and that knowledge may be limited. Meta advocates a "spectrum of openness" in which developers decide for themselves how to release their models. According to Meta, if the U.S. were to restrict the ability of companies to release models as open source, it would undermine U.S. interests and allow other countries to take the lead.

Sources: