Researchers describe how to tell if ChatGPT is confabulating
Finding out whether the AI is uncertain about facts or phrasing is the key.
@arstechnica why are we creating tools which lie to us and provide inaccurate info? I'm supposed to become Sherlock Holmes now?
@donlamb_1
Because that means less expenses, by laying off expensive humans and replacing them with cheaper #AI? IOW, because of increased profits. I blame end-stage capitalism... Also, AIs won't unionize, so that is extra risk reduction there. Really, economy101 stuff. The ensuing all-round #enshittification is just continuing current trends where #profit is only important measurement. The impeding end of civilization is unfortunate, though. But think of the short-term profits!
@arstechnica
@arstechnica so like when it’s trained on Wikipedia?
@arstechnica so the best way to train llms is with highly curated data, not just random web scraping? Who would’ve guessed except anyone with a brain