Must confess that until reading this article I did not had an informed opinion on how to balance the benefits of open sourcing AI models and the risks and perils that such decision could have to the society as a all. Lessing’s opinion seems a balanced one, whereby the 4 key components - i) source code, ii) training data, iii) Inference code & iv) model weights - of a AI should be freely shared, but the first 3 should be shared without any constraint, but the fourth (model weights) should be with significant restrictions as it is there that the potential big risks resides, and by freeing up the first 3 it will foster learning, it will underpin the future development of the technology, creating a vibrant ecosystem of future developers, thus fulfilling the open-source objective. The article also brought to my attention the Hugging Face platform, that i will try to explore in the future. “…Hugging Face, an AI community platform, offers over 350,000 AI and machine-learning models, 75,000 datasets and 150,000 demonstration applications, all open-source and publicly available. ..” Will continue to read on the subject, also to have divergent opinions and try to fine-tune my own, but as mentioned before Lessing’s opinion seems a reasonable and balanced one. “…Not all AI models should be freely available, argues a legal scholar https://www.economist.com/by-invitation/2024/07/29/not-all-ai-models-should-be-freely-available-argues-a-legal-scholar from The Economist…”
- Pedro
Read on Substack
No comments:
Post a Comment