"In my opinion, OSI continues with its flawed 'one size fits all' approach rather than helping to better define the 'spectrum' for open source and AI." Peter Zaitsev, founder of Percona, argues that the new definition lacks nuance and fails to address the complexities of open source AI.
"Training is where the innovation is taking place, the OSI said, so transparency around the code used in training is necessary to allow open source users to study and modify AI systems." The OSI emphasizes the importance of transparency in AI training code for the benefit of open source contributors.
The OSI noted that there are four types of training data – open, public, obtainable, or unshareable – and that all must be shared legally. However, the legal requirements differ, leading some to critique the classification.
Zaitsev takes issue with the OSI's definition, particularly the distinction between 'obtainable' and 'unshareable' data, suggesting it oversimplifies data accessibility issues in open source AI.
Collection
[
|
...
]