Today, artificial intelligence tools are everywhere—whether you’re dedicating an entire AI agent to overhauling your business’s backend, or just using a neat editing feature in Canva that you may not even realize is AI-enabled. In many ways, the presence of the technology in those tasks can remain invisible to the outside world: A ChatGPT-generated Instagram caption isn’t a scandal; nor does it necessarily matter much if the color tones in your renderings were tweaked with an AI photo editing tool.

But if your content was slapped with a label making your AI usage clear, would it change the way you post?

Tech companies and social media platforms are now reckoning with their own responsibility when it comes to helping users determine whether content is original or AI-generated. After all, the stakes are much higher in some cases: Deceptive generated text and images have made it easier to pull off internet scams (like when Taylor Swift appeared to be peddling Le Creuset cookware) or sow disinformation tied to global conflicts.

Offering a degree of transparency in the new AI slop–ridden online landscape, this summer has seen a steady flow of new top-down policies—from legislators, tech giants and social media platforms—that could bring an end to the AI guessing game, and alter some creators’ content strategies along the way. Those efforts include embedded digital watermarks tied to the generative process; automatic labels applied by social media platforms; and new laws requiring user disclosure of AI-generated content.

One such policy—the most overarching of its kind thus far—has been in the works for over two years. In 2023, the European Parliament and European Union member states reached a historic agreement to issue a comprehensive set of laws regulating AI: the EU Artificial Intelligence Act. On August 2 of this year, the transparency rules (Article 50) finally took effect, dictating that text, images, video and audio designed to look authentic must be labelled as AI-generated and contain a digital watermark, The Guardian reports. Companies who fail to comply could face fines of up to 3 percent of their annual revenue. Exemptions apply to users’ personal content, and content that is “evidently artistic,” meaning fictional or satirical.

Closer to home, similar regulations are manifesting on a local level. California became the first state to implement AI-labeling regulations with the passage of the AI Transparency Act, which also took effect August 2. The law requires major tech companies—OpenAI, Anthropic, Google and Microsoft included—to provide users with an AI detection tool, along with issuing disclosures attached to AI-generated or -altered content. Similar regulations are set to roll out in Washington state early next year, while Oregon and New York may be close behind. As Compliance Week reports, however, adherence to the new regulations may take time: An analysis of 13 major tech operators found that more than half have not yet rolled out their detection tools.

Leading the charge in rolling out policies in compliance with the new labeling regulations is Anthropic. This month, the tech giant announced that all models of its Claude AI tool released after August 2—the same day the EU’s transparency code took effect—will automatically include watermarks in generated text, which will carry over even if users copy and paste or edit files. “Watermarking will be applied at the model level, which means it will be present no matter which Claude product or surface the text comes from,” Anthropic’s support page reads. As TechCrunch reports, the company is currently working on extending the same functionality to past models.

Social media platforms like Meta have also committed to following the EU’s code. Instagram users may have already noticed the changes, which the platform began rolling out in 2024, tagging AI-generated content with “Made With AI”—later revised as a broader “AI Info” label to more accurately reflect posts that only utilize minor AI-powered adjustments, such as edits through Photoshop. Just this year, the platform also rolled out a new account label for creators who largely post generated content, allowing them to utilize an “AI creator” badge on their profile.

For some designers, the new features are a welcome addition to the social media landscape. Vancouver-based designer Janie Hungerford points out that an influx of AI-generated interior imagery online provides prospective clients with unreasonable expectations for the types of outcomes design services can provide.

“I think a lot of people are going to be wanting things that maybe can’t even happen because it’s been AI-generated, and then you get into this world of illusion versus reality,” she says. “Kind of like when those HGTV shows came out, and they were like, ‘You can do a kitchen for no money, and it’ll look this beautiful!’ but actually you can’t—or you could, but it probably won’t work, or last very long.”

The trouble is that even when it comes to capturing real spaces, the postproduction process often involves some level of editing, often with the help of AI tools. “I just posted a kitchen on social that I did, and I took the hinges off the door [using an AI tool] because I didn’t like it—it just looked distracting in the photo,” says Hungerford. “But did that change my design? No.”

New York–based designer Sara Mosele—who uses AI to help develop ideas, refine language and explore visual concepts—is open to an AI label on content, but only if the technology played a significant role in the creation process, as opposed to for tasks like editing, brainstorming or refining. The significance, she says, is in identifying AI-generated work from AI-assisted workflows.

“The larger conversation should be about authorship and transparency rather than AI detection,” she says. “In design especially, AI can accelerate exploration, but it doesn’t replace expertise, taste, context or responsibility. The important question is not just ‘Was AI used?’ but ‘What role did it play, and who is ultimately responsible for the work?’”