Advertising always seeks to persuade. But when an "AI-generated" label appears on a banner or a social media post, the buyer often does not understand what exactly it means. The authors of a piece in The Conversation discuss this.
Artificial intelligence (AI) is a program that can generate text, images or video. The label may appear both on advertising where AI only slightly touched up the image and on advertising that is almost entirely made by a machine. Between these extremes there are many variants.
The difference matters: there is evidence that disclosing AI involvement reduces trust. Other studies point to a general consumer resistance, although the reaction depends heavily on what exactly the technology creates.
Disclosure in advertising is not new. Influencers use "#ad", and sponsored material is marked "advertising". But the AI label works differently: it speaks not to why the content is shown but to how it was made. At the same time it does not disclose the roles of the human and the machine.
For decades designers retouched photos and removed backgrounds while retaining control over the concept. AI blurs the line between editing and creation: it can generate elements itself. A person can come up with the idea and refine the result, but the balance of contribution varies greatly.
Because of this, the label provokes judgments: who created the content, how much of it was human, and whether it was appropriate to use AI. This may be why campaigns by Coca-Cola and Gucci made with AI involvement were criticized.
Authorities and platforms are trying to bring order. The European Union wants strict disclosure requirements for creators of generative AI systems. Technologies are emerging that allow digital content, including advertising, to carry information about its origin and edit history.
Details matter. Meta acknowledged that its "Made with AI" label appeared on content with minor AI edits and did not provide enough context. The company replaced the wording with "AI info", and later made the label less noticeable if AI was used for editing rather than creation.
There is also ambiguity in the opposite direction: the labels "human-made" and "no AI" signal human authorship. But do they mean a complete absence of AI or a predominantly human process in which AI played a supporting role? Both types of labels can hide no less than they show.
For now, consumers should think about how they interpret such labels. "AI-generated" says something, but not everything. The label by itself does not mean the result is fake, bad or unreliable. It also does not clarify the role of the human: who came up with the idea, selected and refined the result, checked accuracy and decided what would ultimately come out.
The AI label should be a starting point for evaluating content, not the final one. Understanding how AI was used and how much human judgment is in the result would say far more.