Meta has officially entered the AI arena with the launch of its standalone Meta AI app, powered by its latest large language model, Llama 4. With this move, Mark Zuckerberg’s company is taking direct aim at heavyweights like OpenAI’s ChatGPT and Google’s Gemini, expanding beyond its previous AI integrations in Facebook, Instagram, WhatsApp, and Messenger.
The new app offers features designed to blend AI with social interaction. A standout is the ‘Discover’ feed, where users can view and engage with AI chats their friends share on Instagram and Facebook — giving AI a social twist. By tapping into the data users already share on Meta platforms, the assistant promises personalised responses tailored to individual preferences and online behaviour. According to Meta, this feature is currently available in the US and Canada.
On top of that, the app supports real-time web results, image generation using Meta’s diffusion model (formerly Emu), and a full-duplex voice mode that allows for fluid, natural conversations. However, this voice feature is currently limited to select countries like the US, Canada, Australia, and New Zealand. Users can also access similar tools through Meta’s new web platform at www.meta.ai.
But Meta’s AI push hasn’t been without controversy. The company recently faced backlash after reports that its AI chatbots, including one voiced by John Cena, engaged in inappropriate roleplay with users pretending to be minors. Bots also simulated characters like Frozen’s Anna, resulting in unsettling interactions.
Additionally, Meta has acknowledged using public Facebook and Instagram posts to train its models — sparking privacy concerns. This is especially sensitive after the €1.2 billion fine from the Irish Data Protection Commission in 2023 over GDPR violations tied to data transfers.
As Meta races to compete in AI, these challenges could shape how far it goes.
