Perplexity is in hot water after the BBC accused the AI search startup of quietly scraping its entire archive to train its default model, warning that it’ll head to court if those copies aren’t deleted and a compensation plan isn’t on the table.
In a letter leaked to the Financial Times, the U.K. broadcaster gave Perplexity CEO Aravind Srinivas an ultimatum: stop hoovering up BBC content, erase whatever’s already stored for AI training, and figure out how much you’ll pay for infringing on our IP.
Backed by Nvidia (NASDAQ:NVDA) and Amazon (NASDAQ:AMZN) Co-Founder Jeff BezosPerplexity didn’t hold back in its response, branding the BBC’s move manipulative and opportunistic and suggesting the broadcaster just doesn’t get how web tech and copyright law actually work.
This isn’t their first brush with big publishersback in October, the New York Times sent a cease?and?desist over similar scraping claimsso it underscores a bigger headache for model builders: litigation risk is growing just as fast as the data they’re hungry for. Some of the deep-pocketed players are sidestepping this by licensing content outright, but for cash-strapped startups, that could mean tighter margins and slower rollouts if court dates start piling up.
This article first appeared on GuruFocus.