Chat GPT's Falsehoods

Tech futurists have been saying for decades that artificial intelligence will transform the way we live. In some ways, it already has: Think autocorrect, Siri, facial recognition. But ChatGPT and other generative A.I. models are also prone to getting things wrong—and whether the programs will improve with time is not altogether clear. So what purpose, exactly, does this iteration of A.I. actually serve, how is it likely to be adopted, and who stands to benefit (or suffer) from it? On episode 67 of The Politics of Everything, hosts Laura Marsh and Alex Pareene talk with Washington Post reporter Will Oremus about a troubling tale of A.I. fabulism; with science fiction author Ted Chiang about ramifications of an A.I-polluted internet; and with linguist Emily M. Bender about what large-language models can and cannot do—and whether we’re asking the right questions about this technology.

Learn more about your ad choices. Visit megaphone.fm/adchoices