Automatically applying protections, these accounts limit who can contact them and what content they see, providing parents with peace of mind about the social media interactions of their children.
The models include image-to-text and text-to-music generation models, a multi-token prediction model, and a technique for detecting AI-generated speech.
This chip's architecture is fundamentally focused on providing the right balance of compute, memory bandwidth and memory capacity for serving ranking and recommendation models, says Meta.
"We recognise the concerns around AI-generated misinformation and believe combating this requires concrete and cooperative measures," Meta's Shivnath Thukral says in a statement.