The parent company of Facebook and Instagram is gunning to match Microsoft-backed OpenAI’s most advanced model of ChatGPT, and reportedly plans to start training its large language model in early 2024. Meta wants the new system to be several times more powerful than Llama 2, its current AI language model it released with Microsoft in July. (The Wall Street Journal)

Talking point: Like many big tech companies, Meta faces pressure to either incorporate or keep pace with ChatGPT 4. Companies like Microsoft have chosen to invest in that company, but Meta, laden with major AI starpower among its workforce, seems to be forging its own path. Meta’s AI team includes Yann LeCun, who alongside Yoshua Bengio and Geoffrey Hinton is considered a godfather of deep learning technology, and Montreal-based AI researcher Joelle Pineau. Meta’s AI ambitions could face legal risks, however, if its open source model ventures into copyright-protected information or is used to spread disinformation, the Journal reported. Meta CEO Mark Zuckerberg is among the major tech CEOs expected to meet U.S. Senate Majority Leader Chuck Schumer this week to discuss AI “doomsday scenarios.”