New Meta chatbot with artificial intelligence is trained to work with public posts on Facebook and Instagram

Katie Paul

MENLO PARK, California (Reuters) – Meta Platforms used public posts on Facebook and Instagram to train its new Meta AI virtual assistant, but excluded private posts shared only with family and friends to respect consumer privacy, the company’s chief policy officer said Reuters in an interview.

Meta also did not use private chats on its messaging services as training data for the model and took steps to filter personal data from public datasets used for training, Meta’s president of global affairs Nick Clegg said on the sidelines of the company’s annual Connect conference. This week.

“We tried to eliminate datasets that were dominated by personal information,” Clegg said, adding that the “vast majority” of the data Meta used for training was publicly available.

He cited LinkedIn as an example of a website whose content Meta deliberately chose not to use due to privacy concerns.

Clegg’s comments come as tech companies including Meta, OpenAI and Alphabet’s Google have come under fire for using information gleaned from the internet without permission to train their artificial intelligence models, which ingest huge amounts of data to synthesize information and creating images.

Companies are considering how to handle proprietary or copyrighted material removed during this process that their artificial intelligence systems can reproduce, while also facing lawsuits from creators accusing them of copyright infringement.

Meta AI was the most significant product among the first consumer-facing artificial intelligence tools CEO Mark Zuckerberg unveiled Wednesday at the annual Meta Connect product conference. This year’s event was dominated by talk about artificial intelligence, unlike past conferences that focused on augmented and virtual reality.

Meta built the assistant using a custom model based on the powerful Llama 2 large language model, which the company released for public commercial use in July, the company said.

The story continues

It will be able to generate text, audio and images and will have access to information in real time thanks to a partnership with Microsoft’s Bing search engine.

Public posts on Facebook and Instagram that were used to train Meta AI included both text and photos, Clegg said.

He said Meta has also imposed security restrictions on what content the tool can generate, such as prohibiting the creation of photorealistic images of public figures.

Regarding copyrighted material, Clegg said he expects “a fair amount of litigation” on the question of “whether creative content is subject to the existing fair use doctrine” which allows limited use of protected works for such purposes. like commentary, research, etc. and parody.

“We think so, but I strongly suspect it will lead to litigation,” Clegg said.

Some companies with imaging tools make it easy to reproduce iconic characters like Mickey Mouse, while others pay for the materials or deliberately avoid including them in training data.

OpenAI, for example, this summer signed a six-year deal with content provider Shutterstock to use the company’s libraries of images, videos and music for learning.

When asked whether Meta had taken any similar steps to avoid reproducing copyrighted images, a Meta spokesperson pointed to new terms of service that prohibit users from creating content that violates privacy and intellectual property rights.

(Reporting by Katie Paul in Menlo Park, California; Editing by Kenneth Lee and Matthew Lewis)