The world of typography is facing a new question: when a font is learned by a machine, is it still simply being used, or has it become part of the data?
There was a time when the biggest concerns surrounding a font foundry revolved around illegal distribution, font piracy, or improper license usage. Now, a much more complex issue has emerged: artificial intelligence.
Generative AI has changed the way the creative industry works. Designers use it to find ideas, explore visuals, create moodboards, and even generate images within seconds. At the same time, the same technology has raised questions about how creative works are collected, learned from, and used by AI systems.
For type designers, this issue feels particularly close to home.
A font is not simply an OTF or TTF file that can be installed on a computer. Behind it lies a lengthy process involving research, sketching, character construction, spacing, kerning, testing, revisions, and the development of a consistent character system.
Once a typeface enters the digital world, the question begins to change.
What happens when fonts created for humans to use start being learned by machines?
We have become accustomed to seeing fonts as finished products. We download them, install them, and use them to create logos, posters, packaging, websites, or social media content.
The relationship is relatively simple: type designers create fonts, designers use them, and licenses determine how those fonts may be used.
AI makes this relationship more complicated.
A machine learning system does not see a typeface in the same way a designer does. To a human, a font may feel vintage, editorial, playful, or sophisticated. To a machine, those characteristics can become part of visual patterns processed alongside other data.
This is where an important distinction emerges between using a font and learning from a font.
A designer uses a typeface to create something new. In certain contexts, AI can use visual data as part of a learning process to produce something else.
Both involve a creative work.
But their purposes and processes are very different.
And this distinction is what is beginning to make some creators reconsider how their typographic assets should be protected.
From the outside, creating a font may look like drawing an alphabet. But for a type designer, the actual work goes much further than that. A single character does not exist in isolation. Each letterform needs to work alongside the others. Width, height, spacing, kerning, and visual rhythm all need to be considered as part of a single system. When a font contains hundreds or even thousands of glyphs, the complexity increases even further.
Then there are language requirements. Accented characters. Numbers. Punctuation. Ligatures. Alternate characters. Stylistic sets. Everything needs to maintain a consistent visual relationship. That is why two fonts that look “similar” at first glance may not share the same design system. And that is precisely where the value of a typeface lies.
It’s value is not only in the letterforms visible on screen, but also in the small decisions that build the entire system. That is why when a typographic work is treated merely as a collection of visual data, some type designers feel that something is missing from the conversation.
What is missing is the creative process behind it.
Interestingly, the debate around AI is not really about whether type designers like or dislike technology. Most of the design industry already depends heavily on technology. Font editors are technology. Design software is technology. Variable fonts are technology. OpenType is technology.
Type designers do not exist outside technological development. Quite the opposite they are often among those who adopt new technologies the fastest. The problem emerges when technology develops faster than the rules governing it. A foundry may have no objection to a designer using its font to create a poster that is then sold to a client. That is, after all, one of the purposes for which fonts are created.
But what if that font is used as part of the training process for an AI system?
Does that still count as design use?
Does it require permission?
Is it different if the font is only used to generate an image?
And who should benefit when a commercial system grows by making use of creative work produced by someone else? Not all of these questions have simple answers.
YOU MAY LIKE
Handdrawn Font and Vector Set for Summer 2026 Project
Typography Trends Today Are Moving Toward a Bolder Direction
How to Leverage Figma Community for Your Portfolio
There is another issue that I find even more interesting for the design world. For years, a typeface has been one of the ways a brand or designer can establish a visual identity. A particular serif can immediately feel editorial. A handwritten script can evoke a handmade quality. A condensed display font can create a strong sense of energy. A distressed typeface can bring a sense of heritage. These characteristics take time to build.
AI, however, has made the process of generating visual styles dramatically faster. We can now ask a system to generate visuals with a particular character without necessarily understanding the entire design process that shaped that character. This does not mean that AI automatically “steals” a font. The issue is far more complex than that.
But this development brings us back to an old question in a new context:
How far can a visual style be considered someone’s property?
And when that style is built from the work of many creators, how do we determine the boundary between inspiration, learning, imitation, and exploitation?
This shift in attitude has led some foundries to pay greater attention to protecting their work. The approach can take many forms. Licenses may become more specific. Usage information may be clarified. Font distribution may be monitored more carefully. Some creators are also beginning to consider how their assets can be accessed by automated systems that collect data from the internet.
However, I believe there is one important distinction to understand. Protecting fonts from AI use does not necessarily mean that a foundry is against AI. A type designer can use AI to assist with research or exploration while still disagreeing with their work being used to train a model without permission. These two positions are not contradictory. We can embrace technology while still questioning how that technology uses the work produced by humans. This kind of critical approach may be exactly what the creative industry needs right now.
There is a tendency to view digital assets as something that can easily be duplicated because their final form is simply a file. A font can be sent over the internet within seconds. It is relatively small. It can be stored on multiple devices. It can be used repeatedly. But ease of distribution does not mean that the process of creating it is easy.
Imagine a type designer spending days or even months refining a type family. Hundreds of small decisions may be made before the final version is released. When we look at an OTF file, we only see the final result.
We do not see the hours of work behind it.
We do not see the characters that were redrawn again and again.
We do not see the spacing that was tested repeatedly.
We do not see the decisions behind why one form was kept while another was discarded.
The file is the final result. It is not the entirety of the work.
In my opinion, the question of whether AI will replace type designers is not actually the most interesting one. Type design contains many aspects that are difficult to reduce to simply generating letterforms. There is context, history, function, systems, legibility, and aesthetic decision-making, all working together.
What is more likely to change is the ecosystem in which type designers work.
The way fonts are discovered may change.
The way fonts are marketed may change.
The way creative work is protected may change.
Even the way we define the use of a typeface may change.
And these changes alone are enough to make foundries reconsider their strategies.
Until now, font licenses have generally dealt with relatively straightforward questions.
Can the font be used for a logo?
Can it be used on merchandise?
Can it be installed on a website?
Does it require an app license?
These questions remain relevant.
However, AI technology introduces potential forms of use that were not necessarily anticipated when many font licenses were originally created. That is why conversations around font licensing will likely become increasingly complex. Not because fonts themselves have suddenly changed.
But because the way humans use technology with fonts is changing. This is a pattern that has occurred repeatedly throughout the history of the creative industry. When new technologies emerge, existing rules often need to be reconsidered. AI may simply be the next chapter.
As someone who works with fonts, I see this issue as more than a technological question. It is also about respecting the creative process. When someone purchases a font, they are not simply buying letterforms. They are gaining access to a design system created through someone’s creative process. That is why licensing remains important. That is why illegal distribution is a problem.
And with the emergence of AI, perhaps it is time to start discussing another layer: how typographic work is used as data. I do not think the answer is to isolate ourselves from technology. Instead, the creative industry needs to find ways for technology to develop without diminishing the rights and value of the people who create the work.
Because if technology continues to become more intelligent while creators increasingly lose control over their own work, then something is out of balance within the ecosystem.
NEXT READING
Why do so many designs look similar today?
Optimizing Variable Fonts: Techniques for Maintaining Text Sharpness When Switching to Dark Mode
Professional Designer Secrets: Mastering Mix & Match Retro Font Duos for Global Projects
Ultimately, fonts are just one small example of a much larger shift taking place across the creative industry. AI is forcing us to reconsider what it means to “use” a creative work. In the past, using a font meant typing with it. Today, technology allows a system to process visual works on a much larger scale. The difference may sound technical, but its implications touch on something fundamental to the creative world: ownership, permission, attribution, and the value of the creative process.
Perhaps that is why more foundries are becoming cautious. Not because they are afraid of technology. But because they want to make sure that when technology learns from human work, the humans who created that work do not become the forgotten part of the process. And perhaps this is the conversation we actually need to start. Not about whether AI will take over typography.
But about what we want the future of typography to look like when humans and machines are both learning from the visual language we create.
