Meta’s new AI assistant trained on public Facebook and Instagram posts

Introduction

Meta, the parent company of Facebook and Instagram, has revealed that parts of its new AI virtual assistant were trained on public posts from these platforms. A top Meta policy executive clarified that private posts shared only with family and friends were excluded from the training data to prioritize consumer privacy. Additionally, private chats on Meta's messaging services were not utilized for training the AI model. This move by Meta comes amidst criticism faced by tech companies like Meta, OpenAI, and Google for using unauthorized information from the internet to train their AI models. In another development, Meta's CEO Mark Zuckerberg announced a new generation of Meta's smart glasses, equipped with the AI assistant, which are set to launch in October for $ 299 and can live stream directly to Facebook and Instagram.

Keyword

Meta, AI virtual assistant, Facebook, Instagram, privacy, training data, smart glasses, Mark Zuckerberg, live stream

FAQ

1. What data did Meta use to train its new AI virtual assistant?

Meta used public posts from Facebook and Instagram but excluded private posts shared with family and friends, as well as private chats on their messaging services.

2. How much will Meta's new smart glasses cost and what features will they include?

The new generation of Meta's smart glasses will be priced at $ 299 and will incorporate the new AI assistant, allowing users to live stream what they see directly to Facebook and Instagram.

3. Why have tech companies like Meta faced criticism for their AI models?

Tech companies have been criticized for using information scraped from the internet without authorization to train their AI models, raising concerns about privacy and data usage.