I'm not against ethical AI, I just haven't seen it yet.
@rasterweb
The way I see it there are four major ethical issues with today's AIs:
1. Disproportionately large energy consumption
2. Intellectual property theft
3. Blurring the line between reality and fiction with deep fakes and hallucinations
4. Shoving AI into every fucking thing (which exacerbates the above).
With that in mind, your acquaintance's model seems quite benign. There a lot of cool and interesting stuff you can do with small and specialized models.
Even large and general models (like LLMs) can be built ethically with the above points in mind. The problem (as with most things today) lies with enshitification rather than the technology itself.