The BBC is threatening legal action against US AI startup Perplexity over the alleged unauthorized use of BBC content to train its AI systems. In a letter seen by the Financial Times, the BBC demanded that Perplexity immediately stop scraping its content, delete stored BBC material, and provide financial compensation. The broadcaster says Perplexity copied content verbatim, undermined BBC's own services, and used BBC material to train its standard AI model, Sonar. An internal BBC analysis found that 17 percent of answers generated by Perplexity's chatbot contained significant errors. Perplexity denies the allegations, but the company is already facing lawsuits from other media organizations and is in licensing negotiations with selected publishers.

Ad
Support our independent, free-access reporting. Any contribution helps and secures our future. Support now:
Bank transfer
Matthias is the co-founder and publisher of THE DECODER, exploring how AI is fundamentally changing the relationship between humans and computers.
Join our community
Join the DECODER community on Discord, Reddit or Twitter - we can't wait to meet you.