everyone has an opinion about ai—generated images. almost nobody has examined what they were doing before ai existed. the comparison is overdue.
there is a category of person who will confidently tell you that ai-generated images are a threat to creators, a violation of artistic integrity, and a sign of cultural decay. the same person has a lifestyle blog with forty-seven pinterest images they pulled without permission, without credit, and without a second thought. they have been doing this for eight years.
this is not a defense of ai images. it is a request for consistency.
the conversation about ai and visual culture is real and worth having. but it keeps happening in a context where the old behavior is treated as the neutral baseline and ai is treated as the disruption. that framing is not accurate, and it is worth slowing down to understand why.
pinterest is not a stock photo site. it is a visual discovery engine, meaning almost everything on it was made by someone else, hosted somewhere else, and repinned until the original source became unrecognizable. when you find an image on pinterest and use it on your website, your marketing, your social media, or your mood board, you are almost certainly using copyrighted work.
the copyright on a photograph is automatic. it belongs to the person who took it from the moment the shutter closes, or the moment the graphic was completed. they do not have to register it. they do not have to label it. they do not have to put a watermark on it. the protection exists regardless of whether the image ended up on pinterest through a chain of repins that buried the source entirely.
writing "not my photo" in the caption does not change this. credit is good manners. permission is the law. the warhol v. goldsmith decision in 2023 made this explicit at the supreme court level: acknowledgment does not equal authorization. the court focused on purpose and market impact, not on whether the user was respectful.
the enforcement of this has historically been low. most photographers do not have the time or resources to hunt down every unauthorized use of their work. so the behavior persists, normalized by the fact that almost nobody faces consequences. normalization through unenforcement is not the same as legality. it is the same as jaywalking — widely done, rarely punished, still against the rules, and still carrying risk if the wrong person notices.
this is where the legal situation gets genuinely complicated, in ways the discourse has mostly ignored.
in march 2025, the u.s. court of appeals for the d.c. circuit affirmed that the copyright act requires all eligible work to be authored in the first instance by a human being. what this means practically: a purely ai-generated image — one where you typed a prompt and downloaded the output — almost certainly cannot be copyrighted. nobody owns it. it enters a kind of gray zone between public domain and legal uncertainty.
this is the opposite of the pinterest situation. the pinterest photo has an owner. the ai image, in most cases, does not. when you use a pinterest photo without permission, you are taking something from someone specific. when you use a basic ai-generated image, you are using something that belongs, legally, to no one.
the complication is that the most significant risk with ai-generated images is that they may inadvertently reproduce elements of existing copyrighted works. ai models are trained on billions of images, many protected by copyright. when these systems generate new images, they sometimes produce content that bears striking similarities to images in their training data. so the ai image is not legally clean by default — it just has different potential legal issues than the pinterest photo, and different people bearing the risk.
the legal comparison favors ai images, mostly. but law and culture are not the same thing, and the cultural argument deserves its own examination.
the people who object to ai images on cultural grounds usually argue that ai was trained on human-made work without consent or compensation, and that using ai-generated images therefore participates in a system that extracted from creators. this is a legitimate concern. the lawsuits against stability ai, midjourney, and openai are not frivolous. in may 2025, the copyright office concluded that ai developers who use copyrighted works to train models that generate content competing with original works are going beyond fair use. the harm to creators from ai training data is real and unresolved.
but the person making this cultural argument while running a pinterest-sourced visual identity is making a selective stand. they have decided that the harm done by ai training data is unacceptable while the harm done by simply taking photographers' work for years was acceptable, or at minimum not worth thinking about. the inconsistency is not a minor footnote. it is the whole argument.
the underlying assumption in both cases is the same: creative work exists for your use until someone makes using it inconvenient. pinterest photos were fine because the consequences were rare. ai images are a problem because the conversation made them visible. the moral position in both cases is opportunistic rather than principled.
if you are genuinely concerned about the relationship between creative tools and creators, consistency requires looking at all of it. that means the pinterest behavior, the google images scrape, the stock photo site where you used the free trial and downloaded fifty images, the template you used without reading the commercial license. all of it sits in the same category as the ai training data concern — people using creative work in ways the original creators did not authorize.
the path forward, for anyone who wants to hold a coherent position, is not to pick the newest tool as the villain while excusing everything that came before it. it is to develop a genuine relationship with the question of where visual work comes from and what it costs the people who made it.
practically, this means sourcing images legally — licensed stock, creative commons with the right terms, your own photography, or ai tools trained on licensed content specifically. adobe firefly uses only licensed and public domain images. getty images and shutterstock both have ai generators built on their own catalogs. these are not perfect solutions, but they are honest ones.
the alternative is to keep picking fights with the newest technology while the older extraction continues quietly in the background. that is not a creative ethics position. it is a comfort zone dressed up as one.
the ai image debate is worth having. the concern about training data, about displacement of working illustrators and photographers, about what it means to generate rather than create — these are real questions that do not have clean answers yet.
but the debate cannot be honest if it starts from a baseline where taking photographers' work from pinterest without credit or permission for the last decade was just how things worked.
the tool changed. the assumption underneath it did not. that assumption is the problem worth examining.