Several tech and AI companies, including GitHub and HuggingFace, have written to European policymakers asking them to offer more support for the open-source development of artificial intelligence (AI) models, The Verge reported on July 26.

In the paper, which contains a list of recommendations for the European Parliament, the companies asked for clearer definitions, among other things.

Companies co-signed the paper include Creative Commons, EleutherAI, LAION, and Open Future.

The companies moved to ensure that the AI Act currently being developed by EU policymakers aligns with the best standards for ensuring the development of AI.

The EU is one of the regions taking AI regulation seriously, as a draft of its AI Act is already available. But the draft has already attracted criticisms from several quarters.

Most of the criticism concerns the hyper-focus on the application layer and the lack of concise definitions of AI technology. Unsurprisingly, the companies have suggested clearer definitions.

Other recommendations include distinguishing hobbyists and researchers working on open-source AI models from those profiting commercially, setting requirements for different models, and regulating real-world testing of AI projects.

The Act has banned all forms of real-world AI testing. But the companies believe such an outright ban will only affect research and development. They also want sharing of AI tools on open-source libraries to be free from regulation.

If passed, the Act could set a precedent for laws regulating AI in other territories. This is why the AI companies have been actively lobbying to ensure the rules are not too stringent and would not impede the growth of the sector.

While EU policymakers are already working on laws to regulate AI, regulations in the U.S. are currently in the early stages. The lawmakers want to avoid the current lack of regulatory clarity that continues to plague crypto in dealing with AI.

Top AI companies, including Google, Meta, and OpenAI, recently agreed to implement watermarks on AI-created content as part of voluntary efforts to improve safety.