The global race to create a universally recognized AI-free label
It has never been harder to know whether the things we consume — a text, an image, a film, a song — were born from human hands or from an algorithm. With the explosion of generative artificial intelligence, the line between human creation and synthetic production has blurred in ways few people predicted just a few years ago. And this uncertainty is not just some abstract curiosity. It directly affects the trust we place in what we read, watch, and buy. When that trust wavers, the entire market feels it.
This is exactly the scenario where a global race has gained momentum over the past few months. Organizations from different countries — from startups to nonprofits, along with established companies in the UK, Australia, and the United States — are competing to create a universal label capable of identifying products and services made entirely by humans. Declarations like Proudly Human, Human-made, No A.I, and AI-free are already popping up on films, marketing campaigns, books, and websites around the world.
The core idea is to offer consumers an AI-free certification that works similarly to the Fair Trade label — the one that signals ethical practices throughout the supply chain. The goal is that anyone can recognize the label immediately and trust it without needing to dig deep into the origins of every product.
The problem, as a recent investigation by BBC News revealed, is that at least eight different initiatives are already trying to establish their own label at the same time 🤯. And without a single globally accepted standard, the risk is clear: instead of bringing clarity, so many competing labels could end up creating even more confusion for the person on the other side of the screen or the shelf.
Consumer expert Dr. Amna Khan from Manchester Metropolitan University summed up the situation well: AI is creating significant disruption, and competing definitions of what counts as human-made are confusing consumers. In her view, a universal definition is essential to build trust, clarity, and confidence.
What is driving the demand for certified human products
The logic behind AI-free certification did not come out of nowhere. It follows a broader pushback against the use of generative AI tools that are replacing human labor and creativity across entire industries. Fashion, advertising, publishing, customer service, and music are just a few of the fields that have already felt the direct impact of this large-scale automation.
The market already knows the power of a well-positioned label. Organic seals on food, energy efficiency ratings on appliances, and accessibility badges on websites are all examples of how a simple icon can completely change the perceived value of a product. In the case of human products, the label would function as a guarantee of creative authenticity — a stamp confirming that the content was born from real human hands, minds, and experiences, without interference from language models or image and video generators.
The problem is that the current fragmentation threatens to undermine the very trust these labels are supposed to build. Imagine browsing an online bookstore and finding three books with three different labels, each claiming the work was made exclusively by humans. Which one deserves your trust? The American one, the Australian one, or the British one? Without a standardized set of criteria, consumers are stuck in a limbo where the presence of multiple labels can generate just as much distrust as having no labels at all.
How existing certifications work
Among the initiatives mapped by the BBC, certification models vary quite a bit in terms of rigor and cost. Some labels, like those offered by platforms such as no-ai-icon.com, ai-free.io, and notbyai.fyi, can be downloaded by anyone for free or for a fee, with little to no auditing involved. In other words, the creator downloads the icon and slaps it on their product pretty much on their own.
Receive the best innovation content in your email.
All the news, tips, trends, and resources you're looking for, delivered to your inbox.
By subscribing to the newsletter, you agree to receive communications from Método Viral. We are committed to always protecting and respecting your privacy.
Other systems, like aifreecert, require payment and implement a rigorous verification process to determine whether a given product used artificial intelligence in its creation. These auditors use professional analysts combined with AI detection software to validate the claims.
The Not By AI label, for example, offers both a verification service and the option to simply download and use their badges for a fee, with versions available in different languages. It is a flexible model, but it raises questions about how much the non-audited version actually guarantees anything for the consumer.
For experts in user experience and information design, the effectiveness of any labeling system fundamentally depends on simplicity. If the consumer needs to visit an external website, scan a QR code, and read an audit report to understand what a label means, the system has already failed. The strength of Fair Trade, for example, lies in the fact that decades of consistent presence have created an automatic association between the logo and the idea of ethical trade.
The technical challenge of defining what AI-free means
It sounds simple in theory, but defining what actually counts as human-made is a considerable technical puzzle. AI is so deeply integrated into everyday platforms and tools that drawing a clear line between use and non-use has become genuinely complicated.
AI researcher Sasha Luccioni put the question bluntly: AI is so ubiquitous and so integrated into different platforms and services that it is truly complicated to establish what it means to be AI-free. From a technical standpoint, implementation is difficult. For her, AI is a spectrum, and the market needs more comprehensive certification systems rather than a binary approach of AI versus no AI.
Think about a designer who uses AI to generate variations of a visual concept and then manually refines everything. Or a journalist who uses an AI assistant to transcribe interviews but writes every single line of the final piece on their own. These gray areas make up the majority of real-world situations, and most current certifications still do not know how to handle them fairly and transparently 🧩.
Another important point is how fast the tools are evolving. What clearly looks machine-generated today could become indistinguishable from human work in just a few months. Keeping a certification relevant and accurate at this pace requires continuous investment in research and constant updates to verification methods.
The concept of generative AI-free gaining traction
Faced with the difficulty of establishing an absolute concept of AI-free, some market players have decided to focus specifically on generative AI — meaning the chatbots and models that create text, code, music, or video from human prompts.
A notable example came from the film industry. In the end credits of the thriller Heretic (2024), starring Hugh Grant, the producers included an explicit statement: no generative AI was used in the production of this film. It was a decision that turned heads and paved the way for other initiatives in the industry.
Film distributor The Mise en scène Company took the idea further and recently added a No AI was used label to the poster of their latest film, which was written, directed, and edited largely by a single person. The distributor also published its own online rating system, hoping others in the industry will follow suit.
The company’s CEO, Paul Yates, explained the motivation: they support the AI industry and think it is an exciting time, but they believe that as a result of AI-generated content, there is an economic premium being placed on human-made content — and they want to position themselves right in that space.
The disruption in the arts and publishing industry
The creative industry is by far the field most affected by AI production — and also the epicenter of the resistance. Entire books and full-length films are being produced with generative AI tools much faster and cheaper than through traditional methods.
Bollywood studio Intelliflicks, for example, specializes in producing films entirely with AI and makes a point of openly promoting that fact. On the other end of the spectrum, however, many products that rely heavily on AI simply do not disclose it to the consumer — and that is where the real problem lies.
In the publishing world, British powerhouse Faber and Faber has started placing a Human Written label on some of its books. Author Sarah Hall requested that the label be added to her novel Helm and went further, describing the use of books to train AI models without permission as a kind of creative theft on an industrial scale. However, Faber has not publicly detailed how it classifies its books as human-written or what kind of auditing it performs to ensure no AI was used in the process.
British company Books by People agrees that a reliable standard is needed for how human authorship should be disclosed. Co-founder Esme Dennys pointed out that publishers are navigating a new landscape where books can be produced in minutes instead of months or years, and readers can no longer be sure whether a book reflects a human experience or a machine imitation.
Books by People has already partnered with five publishers and placed its first label on the book Telenova, released in November. Their model charges publishers and requires them to fill out questionnaires about their practices and how they verify their authors. The company also periodically checks book samples to detect AI-generated writing.
The Australian model with reinforced auditing
In Australia, the company Proudly Human uses a similar but even more rigorous system. Its auditors perform checks at every stage of publication, including analysis of any changes made from the original manuscript to the final ebook version. The company is about to announce partnerships with major publishing houses and plans to expand into music, photography, film, and animation.
Founder Alan Finkel argues that systems like his are essential because the industry’s own efforts to analyze and label AI-made content have failed. For him, human-origin certification is necessary, but self-certification is not enough — which is why they implemented a full verification process to ensure the material is genuinely of human origin.
Misinformation as both fuel and consequence
There is an interesting irony in this story. Misinformation itself works simultaneously as the engine and the byproduct of this race for labels. As the engine, because it was precisely the inability to distinguish human content from AI-generated content that created demand for a visual, immediate solution. When deepfakes started circulating as legitimate news footage, and when texts generated by language models began being published as opinion pieces without any disclosure, it became clear that some form of labeling was urgently needed.
But the fragmentation of the label market threatens to turn the solution into part of the problem. Without an international regulatory body or at least a minimum set of standardized criteria, consumers are left in a scenario where the presence of multiple labels can generate just as much distrust as having none at all. Misinformation feeds on itself: the label that was supposed to clarify things ends up raising more questions than answers.
The possible path to a label that actually works
For the idea of certified human products to move from good intentions into the territory of real trust, a few conditions need to be met.
- Independent governance: a label only gains credibility when there is an independent entity behind it, with clear auditing rules, revocation processes for those who fail to meet the criteria, and full transparency about how verifications are conducted.
- Adoption at scale by major platforms: book marketplaces, digital art galleries, music streaming services, and app stores need to integrate the label into their search and recommendation systems for it to gain real visibility.
- Multilingual and multiplatform coverage: a label that only works for text does not solve the problem for people who produce music, illustration, or video. And a label that only covers the English-speaking market does little to help independent creators in Latin America, Africa, or Southeast Asia.
- Flexibility to keep up with technological evolution: definitions need to be adaptable enough to deal with tools that do not even exist yet, without losing the clarity consumers expect.
There is also a cultural dimension that cannot be ignored. In many contexts, the use of generative AI as a supporting tool does not eliminate human authorship — it complements it. Defining what is and what is not a human product will require debates that go far beyond technology, touching on issues of intellectual property, collaborative creativity, and even the very meaning of authorship in the 21st century.
What to expect going forward
At the end of the day, the success of a global AI-free label depends less on verification technology and more on the ability of the organizations involved to cooperate instead of compete. As long as each initiative tries to be the sole winner of this race, the public will remain confused and misinformation about the origins of content will keep growing.
The history of quality labels in other industries shows that consolidation is inevitable — the question is how much time and how much confusion the market will have to endure before getting there. For creators, consumers, and platforms, the ideal scenario is one where a single glance at a label is enough to know: this was made by real people. And that clarity, more than any detection technology, is what will determine whether AI-free certification becomes the next Fair Trade or just another logo lost in a sea of empty promises.
