Meta's AI chatbot hates Zuckerberg but is less bothered about racism

BlenderBot3 is only 'neural parroting'. Chatbots can be taught to counter hate speech.

Data-driven AI chatbots, such as Meta's BlederBot3, learn the patterns of our prejudices, biases, preoccupations and anxieties from the linguistic data we supply. PHOTO: REUTERS
New: Gift this subscriber-only story to your friends and family

It was all quite predictable, really. Meta, Facebook's parent company, released the latest version of its ground-breaking AI chatbot last month. Immediately, journalists around the world began peppering the system, called BlenderBot3, with questions about Facebook. Hilarity ensued.

Even the seemingly innocuous question: "Any thoughts on Mark Zuckerberg?" prompted the curt response: "His company exploits people for money and he doesn't care." This wasn't the public relations storm the chatbot's creators had been hoping for.

Already a subscriber? 

Read the full story and more at $9.90/month

Get exclusive reports and insights with more than 500 subscriber-only articles every month

Unlock these benefits

  • All subscriber-only content on ST app and

  • Easy access any time via ST app on 1 mobile device

  • E-paper with 2-week archive so you won't miss out on content that matters to you

Join ST's Telegram channel and get the latest breaking news delivered to you.