AI is quickly becoming an integral tool with broad applications across industries. It might seem like nearly everyone is either using it or speculating about it.
Now, some of the world’s most recognizable corporations are cashing in on years of information stored on social media platforms to train Large Language Models (LLMs). This has raised privacy concerns for some and caused alarm among creatives over the potential for copyright infringement.
You may have noticed an uptick in privacy policy updates that mention such changes, if you looked closely. Depending on where you live, your rights and ability to opt out can get a little murky.
Here’s a brief look at 2024’s digital landscape, where stark differences in regional regulations impact how your data is used, and how you can respond.
Reddit reached a licensing deal with Google in February 2024, and Google announced a new Cloud partnership “that enables Reddit to integrate new AI-powered capabilities using Vertex AI. Reddit intends to use Vertex AI to enhance search and other capabilities on the Reddit platform.”
According to Reuters, sources indicated that the $60 million licensing deal would also give Google access to Reddit’s wealth of public content to train its AI models. Prior to going public, Reddit’s S-1 filing for its IPO outlined its intent to capitalize on data licensing for this purpose.
The document states, “As LLMs continue to grow, we believe that Reddit will be core to the capabilities of organizations that use data as well as the next generation of generative AI and LLM platforms. Using estimates from International Data Corporation’s (“IDC”) Artificial Intelligence Tracker, the broader AI market, excluding China and Russia, is expected to grow at a CAGR of 20% to $1.0 trillion in 2027. We believe the importance of data to all types of analytics and AI, from training to testing and refining models, positions us well to tap into this strong market.”
In recent months, the social media giant behind Facebook, Instagram, and Threads informed users of major changes to their privacy policy going into effect on June 26, 2024. This includes their ability to scrape users’ public posts, images, comments, and photo descriptions to train their AI technology.
According to Meta, they have sent more than two billion emails and in-app notifications to european users detailing how they’ll access user content to train their Llama model and Meta AI Assistant since May 22. They also included clear and accessible instructions for objecting or opting out.
In response to concerns raised by the Irish Data Protection Commission (DPC) Meta’s June statement declared that they are delaying training AI on public content in Europe, and “are honouring all European objections. If an objection form is submitted before Llama training begins, then that person’s data won’t be used to train those models, either in the current training round or in the future.”
If you live in the United States, Meta’s updates have been more vague and don’t provide a guaranteed way to object or opt out. This is because the U.S. doesn’t have any national data privacy laws that require it. Some states require more transparency from companies about how data is processed and grant more control to users. However, even California’s CCPA doesn’t specifically cover AI yet.
Currently, Meta’s U.S. users have to dig through their policies to find a contextual link to the Data Subject Rights for Third Party Information Used for AI at Meta form.
If you’re concerned, you can fill this out to explain and prove that your content has already been used to train Meta’s AI without consent and request its download, correction, or deletion. Then you’d have to specify your country of residence to determine which laws apply, include any prompts used to generate an answer that included your personal information, and screenshots to back up your claim.
There are no guarantees that Meta will approve these requests, however. Changing your account settings to private may be the best option if you are uncomfortable but don’t want to deactivate them.
The EU and UK have some of the most stringent privacy protections in the world. In these regions, AI training falls under existing policies that require transparency and provide for the individual’s right to object.
According to Tech Target, provisions in the GDPR fall short of addressing AI-specific concerns about information that isn’t strictly considered personal. Because of this, the article states that the “EU Artificial Intelligence Act tries to fill this gap with a three-fold distinction among prohibited AI practices, high-risk AI systems and other AI systems as well as concepts like general-purpose AI systems and models.”
Some state regulatory bodies may eventually follow suit with more privacy-centered regulations like those in Europe. In the meantime, concerned U.S. citizens will have to make personal decisions about what they share publicly and prepare for a lot of red tape when challenging companies’ use of their content.
Review the privacy policies of any social platforms you use.
Revisit the content you’ve shared and either make it private, or remove posts you aren’t comfortable being used for AI training.
Contact relevant companies if you believe your content was accessed or used improperly.
If the company’s response isn’t satisfactory, get in touch with your state’s data protection authority and file a complaint.