The European parliament was told that any rules requiring developers to monitor or control use of their work “could make it impossible to release open-source AI in Europe”.
Here, the politicians didn’t think about open-source AI models, that could be used by any developer, no matter their values.
It’s important to note that because politicians are not trained technicians and never have:
- Learned how neural networks work
- Learned how LLMs work
- Learned any programming language
- Written a single line of code
- Posted software on git platforms
- Cloned and used AI from Hugging Face
They would have never realized how difficult or even impossible is to monitor every use of AI, and that it’s not possible to make arbitrary regulations due to how difficult it is to control capabilities and outputs of any unpredictable black box program.
Update: It has been decided to exclude open-source AI from regulation, which could make it possible to release any type of AI in Europe: