Navigating AI Data Use on Social Media: Your Options to Opt Out

As social media platforms increasingly leverage user-generated content to train artificial intelligence models, concerns about privacy and consent are rising. A recent announcement from Twitch, owned by Amazon, stood out for allowing users to opt out of having their streams and chats utilized for AI training. While this move was welcomed by many, it also highlighted how few platforms offer such straightforward options.

Most social media sites, in fact, make it challenging to understand how user data is employed, particularly in the context of AI. Training generative AI models, like Google’s Gemini, is often intertwined with routine features, making it nearly impossible for users to discern the specifics of data usage without extensive digging through complex privacy policies.

It appears that the majority of platforms do not provide users any means to opt out of AI training. In essence, logging into these services is interpreted as a tacit agreement to allow the use of personal data for training purposes. However, some platforms do offer limited options for those who wish to manage their data more actively.

Bluesky claims not to use AI models on user data, although its decentralized nature means that posts can be scraped by third parties. Users concerned about this may want to limit their posting activity. On the other hand, Facebook has a notoriously broad policy regarding AI training. It collects data from posts, photos, and even third-party sources, while stopping short of using private messages unless shared voluntarily. Unfortunately, outside of EU regulations, there’s no opt-out option for most users, leaving them with little recourse.

Instagram, owned by Meta like Facebook, mirrors these policies. While users can file complaints if their private data appears in AI outputs, the primary method of controlling data use is to delete accounts or make them private. Similarly, LinkedIn, under Microsoft’s umbrella, leverages user-generated content from profiles and interactions for AI training. Here, users can opt out through the settings, but only for data collected on LinkedIn itself.

Reddit presents a more complex situation. While it doesn't train its own AI, the platform has partnerships with OpenAI and Google that allow these companies to utilize Reddit’s data. Currently, users cannot opt out, short of deleting their accounts, as data scraping has been a long-standing practice.

Snapchat stands out for providing a clearer opt-out mechanism. Users can disable the use of public content for AI training through the app’s settings. However, the risk of third-party scraping remains, urging users to be cautious about what they leave public.

Finally, Threads, another platform under Meta, shares similar data use policies with Facebook and Instagram, complicating users' ability to manage their information.

In summary, while some platforms like Twitch are beginning to offer opt-out options, many others continue to operate under opaque policies regarding AI training. Users are largely left without meaningful control over how their data is used, making it essential to stay informed about each platform's practices and available settings.

Source: lifehacker.com