DISTRIBUTIONS
Generative AI continues to court controversy as Gentoo and NetBSD heavily restrict its implementation.
Despite the clear advantages of having AI create carefully crafted code, its use in major projects has been problematic.For instance, in March this year Nvidia found itself on the wrong end of a lawsuit. This centres on copyrighted material allegedly being used to train LLMs in the Megatron library for the tech giant’s NeMo generative AI framework. In December 2023, the New York Times launched its own lawsuit against Microsoft and OpenAI, claiming that the newspaper’s articles had been used to train ChatGPT and other models.
These legal wrangles likely won’t go away any time soon, so it’s unsurprising that back in February Gentoo Council member Michał Górny proposed a ban on using AI to develop the OS.
His justification for this was threefold: first he cited the above-mentioned copyright concerns. He also touched on quality concerns, given that in Górny’s own words, “LLMs are really great at generating plausible-looking bulls**t.” Finally, he pointed out that there are also ethical concerns, given its ownership by large corporations, which is “driving the ensh**tification of the internet”. Górny’s scatalogical post clearly struck a chord with the Gentoo Council – on 14th April, it voted unanimously to ban any contributions to the project made using natural language processing AI tools (see https://wiki.gentoo. org/wiki/Project:Council/AI_policy).
The resolution adds the important caveat: “This motion can be revisited, should a case been (sic) made over such a tool that does not pose copyright, ethical
