Is it legal for companies to train their AI on my photos without asking me?

The short answer is that it's a huge legal gray area right now, and in most places, companies have been doing it without clear permission. Many AI image generators like Midjourney and Stable Diffusion were trained on massive datasets scraped from the public web, which included billions of images — personal photos, artwork, you name it — without asking the creators or the people in them. Whether this is 'legal' depends on where you live and how the images are used. In the EU, the GDPR gives people a stronger right to have their data deleted, and regulators are starting to push back. In the US, companies often lean on 'fair use' arguments, claiming that training an AI is a transformative process, like a student learning from a textbook. A concrete example of the mess this creates: an artist found that a medical image of her face, taken for a hospital study, had been used without consent in a dataset for training facial recognition AI. She had to fight to get it removed. The insight most people miss is that the problem isn't just about your face appearing in a generated image. It's that the AI model learns the underlying patterns of your face, meaning someone could potentially create new, fake images that look convincingly like you. For a broader look at who owns AI-generated content, our legal guide on AI content and copyright issues breaks down the current rules. **Related**: Can I get my personal data removed from AI training sets? | How do I opt out of my images being used for AI training?