Meta’s “AI personas” might launch as soon as next month, according to company insiders.
According to the Financial Times, the launch of an AI chatbot feature that would be integrated within Meta’s platforms, including Facebook and Instagram, is coming soon. The feature would enable users to chat with various AI personalities for search and recommendations, such as one “in the style of a surfer” for travel tips, or another in the style of Abraham Lincoln, said the FT.
Meta has been on kind of a roll lately. The launch of Twitter-competitor Threads amassed over 100 million signups within a week of launching (although active users quickly dropped off.)
Meta also recently launched the latest version of its large language model Llama 2, which is an open-sourced alternative to closed-source LLMs from OpenAI’s GPT-4, and Google’s PaLM 2. The “personas” tool, powered by Meta’s generative AI model could boost engagement on its platforms, and let Meta flex its AI muscles in the race with OpenAI, Google, and Microsoft for generative AI dominance.
In February, Meta CEO Mark Zuckerberg first teased the development of “AI personas” in a Facebook post. Zuckerberg didn’t provide much information at the time, but said a group across multiple teams was working on creating text-based experiences within Messenger and WhatsApp and image-based within Instagram. The personas are designed to “help people in a variety of ways.”
“Helping people” could also be read as boosting engagement and retaining users in the social media landscape where TikTok is king amongst younger generations.
An AI chatbot could mimic a ChatGPT experience on Facebook, Instagram, and WhatsApp, while feeding content to users from Instagram Reels and other types of social media posts. As the FT notes, chatting with AI personas would also provide more personal data from users, which in turn could inform Meta’s targeted advertising efforts. Meta declined to share details with Mashable about the AI feature or timeline of its release.
