Facebook chatbot shares boomers' relationship questions with the world

Yahoo

a day ago

Facebook users are accidentally sharing legal woes, relationship dramas and health problems with the world after failing to realise that a chatbot they were speaking to was making the messages public.
Internet users have publicly disclosed potentially embarrassing information or private personal details in conversations with an artificial intelligence (AI) app built by Meta.
While the messages do not appear to have been meant for the public, dozens of posts have been shared on Meta AI's public 'Discover' feed.
In one post seen by The Telegraph, a user asked the chatbot to write a character reference ahead of a court hearing, giving their full name.
'A character letter for court can be a crucial document,' Meta's chatbot said. 'To help me write a strong letter, can you tell me a bit more.'
The person posting replied: 'I am hoping the court can find some leniency.'
In another, a man appears to be asking for advice choosing between his wife and another woman. Others users shared long, rambling voice notes.
Mark Zuckerberg's company launched its standalone Meta AI app in April. On it, users can speak to the company's chatbot, asking it questions in a manner similar to OpenAI's ChatGPT.
Public sharing of conversations is not turned on by default, and users have to log in and confirm that they want to publish a conversation.
However, many of the posts suggest users are unaware that their conversations have been aired in public. It suggests people may have opted to publish their conversations without fully realising what they were doing.
In a post on X, Justine Moore, a partner at venture capital firm Andreessen Horowitz, said: 'Wild things are happening on Meta's AI app. The feed is almost entirely boomers who seem to have no idea their conversations with the chatbot are posted publicly.'
In other shared conversations, users appeared to confuse Meta AI for a customer service bot, or asked it to provide technical support, such as helping them to log in. One chat begins: 'Dear Instagram Team, I am writing to respectfully request the reactivation of my Instagram account.'
When it launched Meta AI, the tech company said its public feed was intended as a 'place to share and explore how others are using AI'.
It said: 'You can see the best prompts people are sharing, or remix them to make them your own. And as always, you're in control: nothing is shared to your feed unless you choose to post it.'
Technology giants have been aggressively pushing AI features despite fears that the tools are leaving social media filled with so-called AI 'slop' – nonsense images and conversations generated by bots.
AI chatbots have been involved in a series of blunders. A Google chatbot last year told its users it was safe to eat rocks. In 2023, a chatbot from Microsoft went rogue and repeatedly expressed its love for users.
Meta was contacted for comment.
Broaden your horizons with award-winning British journalism. Try The Telegraph free for 1 month with unlimited access to our award-winning website, exclusive app, money-saving offers and more.