Meta will use users’ social activities to train its AI models

Meta announced that it plans to use public posts and images from Facebook and Instagram to train its artificial intelligence tools. Meta has also revealed its new privacy policy, which will be effective June 26, 2024.

Meta recently announced new changes to its privacy policy, including provisions for using user data to train its AI tools. This means that posts, comments, images, image captions, and stories shared publicly will be used in AI training. The company has clarified that data from users who are over 18 years of age will be used for this purpose.

Buy physical gold and silver online

Meta’s new policy will take effect from June 26

Meta will train its AI systems using users’ social activities on Facebook and Instagram. Meta has stated that it will use data as old as 2007 to train and fine-tune its AI tools. It will also include content still present on accounts that have gone dormant and are no longer in use.

Also read: OpenAI secures Reddit content for ChatGPT improvement

By agreeing to terms and conditions on any of Meta’s products, for example, setting up a new account or using an existing account on Facebook or Instagram, users must agree to permission to use their information for training AI tools.

Many users and artists who share their works on their accounts on these platforms are concerned about this. However, Meta has said that it will not use data from private messages (DMs) on Instagram and Facebook, but messages to Meta’s chatbot will be used as training material. Conversations between users and businesses will also be considered for training.

Most users cannot opt-out

Most users don’t have the option to opt-out of allowing their data to train Meta AI, but users in some US states and the EU can opt-out. The process, though, is not very straightforward. Users from the US state of Illinois can opt-out because they have AI protection laws. 

Users from the EU are now also protected by laws like the General Data Protection Regulation (GDPR). Max Schrems, Noyb founder said, that,

“Meta is basically saying that it can use ‘any data from any source for any purpose and make it available to anyone in the world,’ as long as it’s done via ‘AI technology.’ This is clearly the opposite of GDPR compliance. ‘AI technology’ is an extremely broad term.”

For information that users share directly with Meta through Facebook or Instagram, users from Illinois and the EU have the right to object to using their information for training AI. The same goes for information that Meta has obtained from third parties. However, Meta requires its users to fill out forms, provide evidence, and enter a one-time passcode. 

Creators have concerns about AI training

Recently, AI companies have faced lawsuits for using data from newspapers and other publications without permission for AI training. Meta has a history of using and selling users’ data, and now it will use the data for AI training.

Also read: 8 US newspapers drag OpenAI and Microsoft into court for copyright violations

Many artists and creators rely on Instagram to engage with people and share their artwork. The concerning part is that Meta can use their creative works and images to develop its AI tools unless users successfully object to their use. Given the cumbersome process to opt-out, content from many unwilling users might be left at the mercy of Meta to train its AI models, according to DAC’s Head of Legal, Simon Marshall.

Children’s illustrator, Sara Fandrey from Portugal, voiced her concerns after she found out about the new policy updates. One of her videos went viral, in which she gave tips on how to fill out the objection form. Talking to the Australian Broadcasting Corporation (ABC), Sara said,

“These images do not come from thin air. They’re all based on other images. I matter. I am a human being. I have a voice. I have a skill that I have gained that I’ve worked on, and it’s mine.”

Sara pointed to Meta’s AI tools, which can now generate images for users after being trained on images from different sources, including artists’ images. She questioned Meta’s authority to scrap content that is the intellectual property of individual artists to train the company’s AI tools.

Dr. Joanne Gray of the University of Sydney said that the US exemption allows these companies to copy content and create something new using AI. Gray said that these are all new challenges, and how things turn out regarding legal proceedings has to be seen. She mentioned the creators’ concerns and said, “These models copy style. While it’s not copyright infringement to copy style, it is important as an economic asset to people when they become established.”


Cryptopolitan reporting by Aamir Sheikh

About the author

Why invest in physical gold and silver?
文 » A