Hand-drawn portrait versus algorithm: What a hand-drawn portrait can do that an AI-generated image cannot

For over ten years, we have been drawing portraits by hand from photographs. During this time, we have completed over 10,000 commissions, ranging from single portraits to entire family and pet portraits. In our consultations over the last two years, the same question has come up remarkably often: Why even bother with hand-drawn portraits when an image generator can produce a similar image in seconds?

The answer has less to do with nostalgia than one might initially assume, and even less with skepticism towards technology. It lies in what actually happens between the photograph and the finished image, in the decisions made along the way, and in what is ultimately meant to hang on the wall, often for decades to come.

What an algorithm sees and what an artist sees

An image generator works with probabilities. It has been trained on millions of faces and calculated what a plausible face resembling the template might look like. An artist works in the opposite direction. They don't start with probability, but with deviation: the slight asymmetry in a smile, the crease that forms when laughing, the way an ear sticks out slightly. It is precisely by such details that family members recognize a face, not by the statistically most probable version of it.

With a single portrait, this difference is often barely noticeable. However, as is the case in a large portion of our projects, when four or five people from different photos need to be combined into a single image, it becomes clear. An algorithm has to adjust the lighting, perspective, and proportions of the individual images without actually understanding the photos themselves.

At this point, the illustrator makes conscious decisions: Which photograph most closely depicts a person as the family remembers them? This is not technical optimization; it is an interpretation, and it is precisely this interpretation that is later perceived as a likeness.

A difference that is now also reflected in legal terms

What surprises many customers is that the difference between hand-drawn and AI-generated images is no longer just a matter of taste, but increasingly also a legal one. According to current German and European case law, purely AI-generated content is generally not protected by copyright.

Only works based on a recognizable human creative effort are protected, as a detailed explanation on urheberrecht.de shows. For a family portrait intended to be passed down through generations, this is more than just a legal footnote. It means that, in the end, a work is created that actually belongs to someone—namely, the person who drew it.

Why viewers evaluate handmade art differently

The perception of art demonstrably depends on whether viewers know how a picture was created. A study from the University of Hohenheim shows that artworks identified as human-made were consistently rated better in direct comparisons with images labeled AI-generated. The reason lies less in the result itself than in the attributed intention.

A hand-drawn picture visibly embodies the time, concentration, and decisions of a person, even if viewers cannot put them into words. This very knowledge changes how a gift is perceived, especially when it involves memories, such as of grandparents or a deceased pet.

Where the limits of AI lie in family portraits

In practice, the limitations become apparent in three main areas. First, when there are multiple people in one image: a generator often makes faces look too similar, so siblings suddenly appear more alike in the finished picture than they do in real life. Second, with low-quality or old photographic source material, which is often the only available source for memorial portraits of deceased relatives.

An artist can compensate for missing image information with experience and ask specific questions when things are unclear, whereas an algorithm fills in gaps with statistically plausible but invented details. Thirdly, there's the overall approach to the process: an AI-generated image, a hand-drawn portrait, is created through a process in which the family can participate, from selecting the photos to approving the final preview before it's even printed.

This pre-print approval is not a minor detail in practice, but rather the very foundation of trust. Anyone ordering a portrait of a deceased family member wants to see beforehand whether the likeness is accurate and needs the opportunity to make adjustments. An automated process typically doesn't offer this kind of feedback in the same way.

What the actual process in the workshop looks like

A typical commission doesn't begin with a single perfect photograph, but often with several different shots from which a coherent image concept must first emerge. For a portrait of four or five family members, the photos that best complement each other in terms of lighting, gaze, and expression are reviewed first. Only then does the actual drawing begin.

After the initial sketch, the family receives a preview before any printing takes place and can provide feedback if something about the likeness isn't quite right. This kind of feedback loop is rarely provided for with an automatically generated image. A generator delivers a result, not a dialogue.

This is particularly evident in animal portraits, which make up a significant portion of our commissions. A pet's fur, coat pattern, and individual posture often differ more from person to person than is the case with humans. It is precisely these details that determine whether an owner recognizes their own pet in the portrait.

A generator tends to revert to the statistically typical breed image. Meanwhile, an illustrator specifically looks for the one characteristic by which an animal is recognized by its owner, such as a crooked ear or a particular head posture.

What ends up hanging on the wall

Ultimately, it's not the technique behind a picture that determines whether it deserves a place on the wall, but rather what it evokes in the viewer. From over 10,000 commissions at familiesportrait.de, this is the most frequent feedback: people recognize not just a face in a hand-drawn portrait, but a gesture. Someone has taken the time to engage with the photos, asked questions, and made conscious decisions before the first line was drawn.

This cannot be automated, no matter how good the technology may become in the coming years, because it's not a technical problem, but a question of attitude towards the image. For a portrait that is meant to hang on the wall for decades and eventually be passed on to the next generation, this remains the crucial difference, not the resolution and not the speed at which it was created.

Joachim Rodriguez y Romero

Owner and Managing Director of Kunstplaza . Publicist, editor, and passionate blogger in the fields of art, design, and creativity since 2011. Graduated with a degree in web design from university (2008). Further developed creative techniques through courses in freehand drawing, expressive painting, and theatre/acting. Profound knowledge of the art market gained through years of journalistic research and numerous collaborations with key players and institutions in the arts and culture sector.

www.kunstplaza.de