Meta has unveiled a series of celebrity-inspired chatbots, which are available in beta on Facebook Messenger, Instagram, and WhatsApp. The chatbots include a Paris Hilton-based character playing a mystery-solving detective, Snoop Dogg as a dungeon master, and a chatbot modeled after YouTuber Mr. Beast, who is described as “the big brother that roasts you—because he cares.” Meta plans to make the tools used to build these chatbots available for Meta users and businesses to create their own versions in the future. This move sets Meta apart from competitors like Google and OpenAI, who keep their latest AI models proprietary.
The introduction of these chatbots builds on Meta’s open-sourced machine-learning models, which the company has been actively incorporating into its apps and services. The first open-source model, Llama, was released earlier this year, and Meta followed up with the more powerful Llama 2 in July. Meta estimates that these models have been downloaded a total of 30 million times and have given rise to 7,000 derivatives. By integrating open-source AI models into all of its platforms, Meta aims to accelerate its development process. However, not everyone sees Meta’s open-source strategy favorably. Some critics argue that Meta should refrain from distributing the most detailed versions of the Llama model, as it may violate an AI risk-management framework from the US National Institute of Standards and Technology.
Meta’s foray into AI assistants is not new, as the company previously introduced a virtual assistant called M in 2015. However, the majority of M’s responses came from human workers rather than algorithms, and the project was quietly shut down in 2018. With the launch of these celebrity chatbots, Meta aims to offer users unique and engaging interactions while also leveraging the power of open-source AI models to drive innovation.