None Of The AI Text Detector Tools Work

About two weeks ago, there was an avalanche of scary and weird text-based ChatGPT and Bing screenshots online. One post showing a creepy story by Bing’s chatbot about becoming sentient and going rogue has over 1,000 upvotes. A post of Bing getting “jealous of a second Bing” and having “a meltdown begging [the prompter] not to leave or offer a chance at humanity to the other Bing” has around 2,800 upvotes and over 1,100 comments. A tweet of ChatGPT appearing to say that OpenAI programmed it to have a liberal bias has over 620,000 views. And last Thursday, the New York Times published a 10,000-word chat with Bing’s chatbot, nearly devoid of context, whose author tweeted that it “tried to break up [his] marriage,” and that the ‘conversation’, which caused him to have “trouble sleeping,” was “one of the strangest experiences of [his] life.” The tweet currently has 4.3 million views.