Meta could release AI chatbot as soon as September

The reports arrive amid Meta's struggle to improve user retention.
Written by Maria Diaz, Staff Writer
PARIS, FRANCE - JUNE 14: The Meta logo is displayed during the Viva Technology conference at Parc des Expositions Porte de Versailles on June 14, 2023 in Paris, France. Viva Technology, the biggest tech show in Europe but also in a unique digital format, for 4 days of reconnection and relaunch thanks to innovation. The event brings together startups, CEOs, investors, tech leaders and all of the digital transformation players who are shaping the future of the Internet. The annual technology conference, also known as VivaTech, was founded in 2016 by Publicis Groupe and Groupe Les Echos and is dedicated to promoting innovation and startups. (Photo by Chesnot/Getty Images)
Chesnot/Getty Images

Meta is preparing a series of AI-powered chatbots capable of displaying various personas, according to a report in the Financial Times. The different personalities reportedly will include a chatbot that imitates Abraham Lincoln and one that impersonates a surfer.

Meta, the parent company behind Facebook, Instagram, and the recently launched Threads platform, reportedly is releasing these chatbots in an attempt to improve user retention. While Threads became the fastest-growing app ever, the social platform is now losing over half its users.  

Also: Google Assistant is about to get supercharged by generative AI, says new report

Meta's chatbots would potentially compete with ChatGPT, Bing, and Bard and could be released as soon as September.

Last fall, OpenAI released ChatGPT, a chatbot that triggered a generative AI boom across the tech world. Microsoft and Google released their own AI chatbots over the following months, and many other AI startups have since done the same. Apple is also reported to be working on its own chatbot. 

The news of the Meta-backed AI chatbots arrives as Meta strives to maintain user engagement on Threads, a social media app linked to Instagram that launched last month in competition with Twitter.

Meta recently released a commercial version of Llama 2, its proprietary, open-source large language model (LLM), after rumors of an impending widespread launch. Llama 2 is a collection of LLMs in three sizes: 7, 13, and 70 billion parameters; the larger the parameter, the more powerful the model. 

Also: Generative AI and the fourth why: Building trust with your customer 

An LLM works a lot like the brain of an AI chatbot. It's an AI algorithm that collects massive amounts of data to process through deep learning techniques, and the number of parameters typically categorizes it by size. These parameters are measured by weights and biases, which define the strength of connections between words and definitions and the rules the model has to follow to use the words in context.

With the basis already built, Meta seems poised to launch a successful set of AI chatbots in a matter of months. 

Editorial standards