ChatGPT creates phisher's paradise by serving the wrong URLs for major companies

This is, in turn, making the world of comment and forum spam much worse. Site operators could tag all user-submitted links as "nofollow," making their sites useless for SEO spammers. But spammers have learned that most LLM content scraper bots don't care about "nofollow," so they're back to spamming everywhere.

I’m not sure if even for traditional search engines “nofollow” means that the scraper doesn’t follow the link, or that it just does not include it in the PageRank or whatever graph but still uses it for to discover new pages. (Of course, LLMs are far too impenetrable for such a middle ground to exist.)

It reminds me of non-radioactive steel, the kind you can only get from ships sunk before the atomic bomb. Someday, we’ll be scavenging for clean data the same way: pre-AI, uncontaminated by the AI explosion of junk.