The Open Washing AI definition is out

The Open Source AI Definition (OSAID) outlines the principles and requirements for what constitutes Open Source Artificial Intelligence. It emphasizes the need for autonomy, transparency, and collaborative improvement in AI systems, mirroring the benefits seen in traditional open-source software. To qualify as Open Source, an AI system must allow users to utilize, study, modify, and share the system freely. This includes providing access to essential components such as data information, source code, and model parameters. The definition specifies that sufficient data information must be disclosed, including the provenance and characteristics of training data, as well as the complete source code used in the system's development. Additionally, the model parameters must be made available under terms that ensure their accessibility. The document also clarifies that the legal mechanisms for ensuring the freedom of these elements may evolve over time. Overall, the OSAID aims to foster an environment where AI development is open and collaborative, benefiting developers and users alike.

- The Open Source AI Definition promotes autonomy, transparency, and collaborative improvement in AI systems.

- AI systems must allow users to use, study, modify, and share them freely.

- Essential components include data information, source code, and model parameters.

- The definition emphasizes the need for detailed disclosure of training data and its provenance.

- Legal mechanisms for ensuring the freedom of AI elements may evolve over time.