It's similar in that regard that it might be wrong sometimes. It's also similar, as coding LLM may be used to translate code between languages.

That's precisely my point. What does "AI-backed" mean exactly? The RFC doesn't ban LLM, but "AI-backed" contributions. Technically from my understanding and interpretation DeepL would fall under that category and would be banned from being used but in the linked thread people seem to agree that DeepL usage is fine. So where do you draw the line? How do you define which "AI-backed" contributions are still fine and which aren't. Eventually different contributors will have different opinions and it will get messy.

At this points it’s just playing word games imho.

The RFC specifically mentions tools like Bard, ChatGPT and Co-Pilot as what it considers to be ban-worthy.

If I’m going to be extremely pedantic about it: with Co-Pilot, the a actual contents of the contribution are AI generated. With a translator like DeepL, the contents are authored by the contributor, the translation tool just translations what’s already there

Both Deepl and LLM use transformer decoders, a type of generative models.