As AI-generated slop takes over rising swathes of the user-generated Internet because of the rise of enormous language fashions (LLMs) like OpenAI’s GPT, spare a thought for Wikipedia editors. In addition to their typical job of grubbing out unhealthy human edits, they’re having to spend an rising proportion of their time attempting to weed out AI filler.
404 Media has talked to Ilyas Lebleu, an editor on the crowdsourced encyclopedia, who was concerned in founding the “WikiProject AI Cleanup” challenge. The group is attempting to give you greatest practices to detect machine-generated contributions. (And no, earlier than you ask, AI is ineffective for this.)
A specific drawback with AI-generated content material on this context is that it’s virtually at all times improperly sourced. The skill of LLMs to immediately produce reams of plausible-sounding textual content has even led to entire faux entries being uploaded in a bid to sneak hoaxes previous Wikipedia’s human specialists.
