“Art-Official” Intelligence: The Concerns of GenAI in Government
- Danielle Delgado
- Jul 22
- 4 min read
OPINION
Whether you like it or not, the AI revolution is here to stay. It has become nearly impossible to go a day online without artificial intelligence playing some role in our digital experiences, from Google’s AI Overview appearing at the top of search results to the content that fills our social media feeds. As AI continues to change how we consume, create, and interact with information, communities and organizations are steadily contending with both its opportunities and challenges.
One area of the AI boom that has received far less attention is the growing use of AI-generated images in government communications and marketing. While artificial intelligence can offer a quick and inexpensive way to create promotional materials, its use by public institutions raises key questions about environmental impact and public trust.
The City of Laurel recently began incorporating AI imagery into its social media communications, starting in March, when the city’s Facebook page posted a flyer reminding residents to keep their trash container lids closed. I immediately noticed several telltale signs that the image had been generated using artificial intelligence. The cartoonish artwork stood out from the format of previous city communications, some text was misaligned, and the visual message implied that recycling was better than trash.
What ultimately confirmed my suspicions, however, was the city’s seal. Although AI image generators have come a long way since they first became publicly available, they still often struggle to reproduce text, logos, and official symbols. Each quadrant of the seal was incorrect—the knight and horse flanking the emblem were distorted beyond recognition, the banner at the bottom was the wrong shape, and its text was illegible.
Since then, several additional flyers containing AI-generated imagery or AI-generated elements have been published on the city’s social media accounts. Flyers promoting Financial Aid Night, the Juneteenth celebration, and information about the FY2027 budget all contained incorrectly generated versions of the city seal, among other visual oddities. The latter two were eventually removed due to the inaccuracies. While an interim Juneteenth flyer was posted to replace the previous, a revised version was later published that contained a different problem: an incorrectly generated Juneteenth flag. In the official flag, a centered white star surrounded by a bursting outline signifies freedom for African Americans and the emancipation of enslaved people. The curved arc separating the red and blue represents a new horizon and the promise of opportunity. In the AI-generated version, the star is noticeably off-center, and the arc was omitted entirely.
The symbols of our cultural and civic institutions matter, down to the finest detail, because they are carefully designed to convey specific meanings. When those symbols are altered, intentionally or not, part of that meaning is lost. The problem with generative AI is that these models lack true comprehension. They do not actually understand symbolism, history, or culture. They do not know why a star is centered, why a particular shape appears on a seal, or what a design element represents in the abstract. Instead, they generate images by predicting patterns from vast collections of training data, producing what looks visually plausible rather than what is historically or culturally accurate.
Beyond concerns about accuracy and representation, the growing use of AI-generated imagery also raises ethical dilemmas regarding the artists whose work was plagiarized to make these systems possible. All AI image generators were trained using vast collections of images gathered from across the internet. In most cases, the creators of such art were neither asked for permission nor compensated for the use of their work in training AI systems. Artists have attempted to resist the appropriation of their creativity; however, the underlying legal questions surrounding AI-generated art remain unresolved.
These disputes highlight an uncomfortable reality for public institutions. How can a government claim to support artists, creatives, and local culture while relying on technology whose development is built upon the unauthorized use of countless creative works? Even if the use of AI-produced images is ultimately found to be legal, legality alone does not answer the larger question of whether it is the most ethical choice, especially when the environmental impact is also in question.
Data centers have been a hot topic in Maryland politics for good reason. The excessive consumption of water and electricity has strained public infrastructure and outraged environmental advocacy groups. With the ever-increasing integration of AI systems and tools, these data centers are being built to meet the demand. The United Nations University Institute for Water, Environment, and Health in June 20216 published its finding that AI data centers can withdraw millions of liters of water per day as system coolant. Rare minerals are being extracted at unprecedented rates to meet demand for new batteries and hardware, causing further environmental harm and directly contributing to slavery and child labor in the mining industry. So, while many in our community fight to have data centers outside of our respective counties, the more we use AI, the more data centers will need to be created. If not here, in some other jurisdiction—but we all share the same finite resources.
I believe that each person has the right to choose whether or not to use AI. It doesn’t make you a bad person if you use it to check your outgoing emails or to brainstorm ideas, just as you’re not a bad person for indulging in some chocolate now and then (even if both have slavery somewhere in their supply chain). It’s wholly different when our government, the institution that we elect to represent our best interests, chooses to utilize such tools on our behalf. I am hopeful that more sustainable and ethical solutions to the broader problems of AI will be explored, but as it stands, we are nowhere close to that reality. Until then, I believe our municipalities should refrain from integrating generative AI into publications and artwork, speaking in the voice of the people rather than an artificial facsimile.
Danielle Delgado holds a Master of Science in Communications from Purdue University and is an educator and community advocate. In June 2026, she was elected to represent District 21 At-Large on the Prince George’s County Democratic Central Committee.
Comments