How to stop your data from being used for AI training

Social media platforms are using your data to train their AI models, often without telling you

Consumer alerts

Tips & Guides


Updated

Take Action

From social media to search engines, AI now seems to be everywhere. Increasingly, social media companies including Meta, X/Twitter and LinkedIn are using our data to train their AI models, often without our permission.

There are lots of reasons to find this onslaught of AI troubling. Anytime personal data is moved or shared, it increases the risk of data security issues. There’s also environmental impact questions. AI requires a significant amount of power to function. The surge in electricity demand from the data centers that power AI is making the transition to clean energy more difficult.

No matter what your reason is for wanting to keep your data out of AI models, here’s how you can take back some control:

Instagram

Instagram’s parent company Meta has been scraping data from public posts to train and improve its Generative AI models. Meta’s AI assistant is a chatbot that can help draft posts or edit images and is powered by Meta’s newest large language model LLaMA3.

Recently, Meta confirmed that all text and photos publicly posted by adult Facebook and Instagram users since 2007 have been scraped and fed into the company’s artificial intelligence models.

And while Meta claims to not scrape data from users under 18, it doesn’t mean minors’ data hasn’t been ingested by the company’s model. At an Australian government inquiry meeting, Meta’s global privacy director could not clarify whether Meta had scraped data from posts by adult users that were made before they turned 18. And if adults posted pictures of their children on a public account, those photos and data were in fact scraped and used for AI training.

How to opt out of Instagram using your data for AI training:

Unfortunately, Meta does not provide U.S. consumers with an option to opt out of having your account data used to train AI models. 

The only real solution (outside of deleting your posts) is to make your Instagram account private. Meta has said it will not use data/information from private accounts. However, making your Instagram account private only affects future AI training and does not change the fact your data has likely already been collected and used for Meta’s AI model. Right now, there’s unfortunately nothing you can do to change that.

How to make your Instagram account private:

  1. Go to your profile page. Click the icon with three horizontal lines in the corner of your screen.

    Photo by TPIN Staff | TPIN

  2. Scroll down and click Account privacy.

    Photo by TPIN Staff | TPIN

  3. Toggle the Private account switch on.

    Photo by TPIN Staff | TPIN

 

Facebook

Like Instagram, Facebook’s parent company is Meta, which has been scraping data from public posts to train and improve its Generative AI models. The Meta AI chatbot can help draft posts or edit images and is powered by Meta’s newest large language model LLaMA3.

Recently, Meta confirmed that all text and photos publicly posted by adult Facebook and Instagram users since 2007 have been scraped and fed into the company’s artificial intelligence models.

And while Meta claims to not scrape data from users under 18, it doesn’t mean minors’ data hasn’t been ingested by the company’s model. At an Australian government inquiry meeting, Meta’s global privacy director could not clarify whether Meta had scraped data from posts by adult users that were made before they turned 18. And if adults posted pictures of their children on a public account, those photos and data were in fact scraped and used for AI training.

How to opt out of Facebook using your data for AI training:

Unfortunately, Meta does not provide U.S. consumers with an option to opt out of having your account data used to train AI models. 

The only real solution (outside of deleting your posts) is to make your Facebook account private. Meta has said it will not use data/information from private accounts. However, making your Facebook account private only affects future AI training and does not change the fact your data has likely already been collected and used for Meta’s AI model. Right now, there’s unfortunately nothing you can do to change that.

How to make your Facebook account private:

Unfortunately there’s not one setting you can turn on that will make your account private. You have to take a series of steps to limit who can see what parts of your profile. Let’s walk through the steps that will – according to Facebook – stop your future data from being collected for its AI training.

  1. Click your profile icon in the top right corner of your screen and select Settings & privacy.

    Photo by TPIN Staff | TPIN

  2. Select Settings.

    Photo by TPIN Staff | TPIN

  3. In the left hand menu, scroll down until you reach Audience and visibility. Then select Posts.

    Photo by TPIN Staff | TPIN

  4. Click the button to the right of Who can see your future posts? It probably says Public – this is Facebook’s current default.

    Photo by TPIN Staff | TPIN

  5. Select the “audience” you want. Selecting “Friends” means your posts will no longer be public. Click Done. (Feel free to take advantage of the other options Facebook makes available here, like curating a list of Close Friends.)

    Photo by TPIN Staff | TPIN

  6. While you’re in this menu, you can also choose to limit who can see your past posts (though this will not stop the fact Facebook has already used your past data for training). We recommend toggling off the other two options available in this menu – “Allow comment summaries on your posts” and “Allow visual search on your posts.”

    Photo by TPIN Staff | TPIN

X/Twitter

X (formerly Twitter) revealed new terms of service that went into effect on November 15th, which essentially say that by continuing to post on X, users automatically consent to X using their data to train its AI models. This includes for Grok, a chatbot for X users that has gotten attention for spreading election misinformation and allowing users to generate fake images of political figures.

You are currently able to opt out of your public data being used for AI training. However, opting out only applies to the future and does not affect AI training that has already taken place using your data.

How to opt-out of X using your data for AI training:

  1. Click More.

    Photo by TPIN Staff | TPIN

  2. Select Settings and Privacy.

    Photo by TPIN Staff | TPIN

  3. In settings, click Privacy and safety.

    Photo by TPIN Staff | TPIN

  4. Scroll down and select Grok & Third-party Collaborators.

    Photo by TPIN Staff | TPIN

  5. Make sure the box is unchecked.

    Photo by TPIN Staff | TPIN

LinkedIn

In September, it was discovered that LinkedIn had automatically opted its users into allowing their data to be used to train generative AI models. The platform has several AI features intended to assist with hiring, job searching and resume building. LinkedIn made this change to data harvesting without informing its users and before even updating their privacy policy.

Fortunately, it is easy to opt out of sharing your data for AI training. However, opting out only applies to the future and does not affect AI training that has already taken place using your data.

How to opt-out of LinkedIn using your data for AI training:

  1. Go to Me in the top menu bar and then select Settings & Privacy.

    Photo by TPIN Staff | TPIN

  2. Click Data privacy.

    Photo by TPIN Staff | TPIN

  3. Select Data for Generative AI Improvement.

    Photo by TPIN Staff | TPIN

  4. Toggle the Use my data switch to off.

    Photo by TPIN Staff | TPIN

TikTok

Tiktok has not revealed whether or not it uses user data for training its generative AI models. However, the platform has rolled out several generative AI features, such as AI avatars of real people, intended to help brands and content creators with marketing.

Given how other major social media platforms have shifted to harvesting user data to train AI models, it’s possible that TikTok will follow this trend.

We need laws that rein in AI data harvesting

This practice of hoovering up user data for AI training is absurd – and not what you asked for when you made your social account.

Right now, there are no clear laws in the U.S. stopping social media companies from using your data to train their AI models, and no laws requiring them to get your permission first or give you the option to opt out. 

European consumers were sent a notice about Meta’s decision to harvest user data for AI training and given the option to opt out of such data collection. That’s because of the European Union’s General Data Protection Regulation (GDPR), which required Meta to provide such notice and opt-out options. The U.S. has no such national privacy law.

We need strong, comprehensive data privacy laws that establish consumer rights and curb excessive data collection. Companies should only collect and use the data they need to provide the service you are expecting to get.

It would be great to have a strong, comprehensive data privacy law at the federal level. However, in the absence of national protections, states have the opportunity to take the lead.

What else can I do to protect my data?

If you want to ensure that your data is as protected as possible, there are other steps you can take. We’ve got more simple ways you can boost your data security here

See below for even more tips to put you more in control of your information online.

Topics
Authors

Ellen Hengesbach

Don't Sell My Data Campaign, Associate, U.S. PIRG Education Fund

Ellen works on data privacy issues for PIRG's Don't Sell My Data campaign. Ellen lives in Chicago, where she enjoys reading, listening to podcasts and spending time with friends.