The New Digital Gold Rush
Artificial intelligence, particularly generative AI that creates images and text, needs to learn from existing examples. To do this, tech companies build massive datasets by scraping, or automatically collecting, vast quantities of data from the public
internet. This includes text, but more importantly, the billions of images we’ve all shared on platforms like Instagram, Flickr, and personal blogs. These photos are digital gold for AI developers. The more data a model is trained on, the better it becomes at understanding and generating new, complex content. The goal is to build powerful, versatile AI products, but the method is raising serious questions.
The Great Consent Debate
The core of the debate is this: just because a photo is publicly visible, does that mean you’ve consented to it being used to train a commercial AI? Most users upload photos to share moments with friends and followers, not to provide free labour for a multi-billion dollar industry. Tech companies often argue that using publicly available data falls under “fair use,” a legal doctrine that permits limited use of copyrighted material without permission for purposes like research. However, creators, artists, and privacy advocates disagree, arguing that training a commercial product is not fair use and that their work is being repurposed without permission, credit, or compensation. This tension has led to numerous high-profile lawsuits against major AI companies. A recent controversy involving Meta's AI image generator, which could use public Instagram photos in its creations, saw such a backlash that the company quickly removed the feature.
What is AI Literacy?
This is where AI literacy becomes a crucial skill for everyone, not just tech experts. AI literacy is the ability to understand, evaluate, and responsibly use AI technologies. In the context of your data, it means understanding that your digital footprint has value and is being used in ways you might not expect. It's not about learning to code; it’s about grasping the basics of how these systems work, recognising their limitations and biases, and knowing how to protect your personal information. Being AI literate empowers you to ask critical questions about the platforms you use and make informed choices about the data you share.
Your Digital Privacy Toolkit
While the legal battles play out, you have tools at your disposal to exert more control over your images. The most effective step is managing your privacy settings. Making your social media accounts private is the strongest defense, as AI scrapers generally do not access data behind a login or privacy wall. Most major platforms, including Instagram, Facebook, and X (formerly Twitter), have settings that allow you to manage how your data is used for AI training, though many are opt-out, meaning you are automatically included unless you take action. For those who publish work publicly, other methods can help. Adding copyright information to a photo's metadata can serve as a digital fingerprint. Some artists are also using tools like Glaze and Nightshade, which subtly alter images to disrupt AI training models. While no method is foolproof, combining these strategies makes a significant difference.
