The researchers also highlighted the lack of transparency about the data used to train the models. The upcoming European Union’s Artificial Intelligence Act, which will apply less strict regulations to models classed as open, is likely to increase the importance of defining open source in AI. The researchers argue that full openness is crucial for scientific reproducibility, innovation, and accountability.
Key takeaways:
- Many technology giants, including Meta and Microsoft, are claiming their artificial intelligence (AI) models to be 'open source' without disclosing important information about the underlying technology.
- The European Union’s Artificial Intelligence Act will apply less strict regulations to models that are classed as open, making the definition of 'open source' increasingly important.
- A study by Mark Dingemanse and Andreas Liesenfeld found that many models claiming to be open or open source are merely 'open weight', meaning outside researchers can use the models but cannot inspect or customize them.
- Openness in AI models is crucial for reproducibility and innovation in science, and for ensuring that inappropriate or copyrighted data has not been used in the models.