Asymmetrical Benefits Of Artificial Intelligence {Content}

I have a lot of concerns about the usage of AI, even when relying SelfHostingArtificialIntelligence and despite proposing some good pratices Analysis.AgainstPoorArtificialIntelligencePractices. It's mostly economical (concentration of power), ecological (data center usage of energy and water), epistemological (slop and the Web, dead Internet theory) and social (data labeling, annotation, preparation of large datasets from unskilled labor).

Yet, despite how visible all this is my growing concern is not about those problems but about actual usage. There is an assymetry of usage that leads to an assymetry of benefits. Even while ignoring the

legitimate usage

• re-writing own content with disclosure of usage
• brainstorming

illegitimate usage

• scam
• spam
• low quality content with life consequences (e.g. published book on mushroom picking, medical advice)
• addictive content with insufficient guardrails (e.g virtual companions or psychological advice)
• cheating in written examinations
• software re-writing to bypass licensing (fake clean room implementation)

...

References

• https://www.techpolicy.press/how-to-regulate-deepfake-financial-fraud/
• https://www.arte.tv/fr/videos/122187-000-A/l-ia-va-t-elle-tuer-internet/
• https://stopsloppypasta.ai
• (sarcasm) https://malus.sh
• https://github.com/alvi-se/ai-ublock-blacklist
• "How AI Destroys Institutions" https://papers.ssrn.com/sol3/papers.cfm?abstract_id=5870623
  • via https://www.techpolicy.press/considering-how-ai-destroys-democratic-institutions/
• a16z startup https://doublespeed.ai
• bot observatory on fediverse https://about.iftas.org