Sarah Andersen on why AI style copying feels personal

If you have spent any time on the internet over the past decade, you have almost certainly seen Sarah Andersen's work. Her semi-autobiographical comic, Sarah’s Scribbles, turned relatable young-adult anxiety into a global viral hit.

Now, the very simplicity that made her comics beloved has turned them into prime fuel for algorithmic exploitation.

In a personal essay published in The New York Times, Andersen opened up about how her distinctive style survived years of alt-right internet harassment, only to face a far more insidious and automated threat: generative AI.

From 4chan trolls to an automated soul harvest

Back in 2016, Andersen became a primary target on 4chan’s /pol/ board. Because she was a woman drawing simple, feminist-leaning cartoons, trolls found it remarkably easy to hijack her work.

Her characters were edited into violent, neo-Nazi imagery loaded with swastikas and Holocaust denial. Someone even went as far as turning her handwriting into a downloadable typeface, effectively automating her personal visual signature.

Back then, Andersen managed to outpace the harassment because human trolls still had to sit down and physically edit the files. Time and labor were on her side.

That changed in October, when a fan sent her an image on Twitter. The user had simply typed Andersen’s name into an AI generator as a prompt, and the machine spat out an eerie, near-perfect imitation of her style within seconds.

Stripping the human out of human art

For Andersen, the emotional hit was immediate and profound. Her style was not just a visual filter; it was the accumulation of a lifetime of reading comics like Calvin and Hobbes and making deliberate personal choices.

Even when another human artist tries to copy a style, variations in line weight and technique inevitably show through. A generative model, by contrast, digests decades of personal dedication and spits out an imitation in seconds without breaking a sweat.

Why it matters

For artists and creators, this marks a chilling shift in how personal intellectual property is treated online. It is no longer just about meme theft or bad-faith edits. Now, an artist’s distinct visual identity can be effectively automated and mass-produced by anyone with an internet connection.

For the AI industry, Andersen's story puts a harsh spotlight on the ethical debt baked into current training pipelines. Scraping millions of living artists' portfolios to let strangers generate work under their actual names looks increasingly indefensible.

We think this tension will define the next wave of copyright fights. When tools make plagiarism friction-free, they risk destroying the very creative incentives that produced the training data in the first place.