A few readers have expressed concern or chagrin over our use of generative AI here the blog, especially when it comes to images. As I have said in several previous experimental posts utilizing AI in a transparent manner: when we use AI in the creation of posts, we will endeavor to be fully transparent about that useage.
We will continue to use AI imagery to illustrate some of our posts because doing so has overwhelming economic and legal advantages for an organization like the BlogForArizona Grassroots Media Network. We will not always specifically note when we use generative AI for illustrative images because it is generally EXTREMELY obvious to now-accustomed eyes that an image is generated by AI (otherwise, some readers wouldn’t complain about the use of such imagery were it not obvious!). However, if we use realistic photographic imagery that might be mistaken for a real photo, we will disclose that said image is AI and not a photograph, to avoid creating misleading content.
I’ll allow AI (Perplexity, in this case) itself (mostly – parentheticals and bold font are MDB) to explain the advantages…
Political blogs (just like this one!) thrive on timely, visually engaging content that translates complex issues into accessible narratives. Yet, finding appropriate imagery for such content can be one of the most challenging and expensive parts of publication. AI–generated images offer an efficient, ethical, and cost-effective alternative to traditional stock photography or licensed artwork. Their low barriers to use, negligible cost, and complete freedom from copyright entanglements make them especially valuable for independent or small-scale political writers.
Cost and accessibility are primary concerns for most political blogs, many of which operate on minimal budgets or serve as passion projects rather than revenue-generating platforms (like BlogForArizona, which keeps the lights on entirely due to reader donations and the small but relatively steady revenue Google AdSense generates). Commissioning custom artwork or photographs can be prohibitively expensive, and subscription-based stock image services quickly add up in price. In contrast, AI image generators have made professional-quality visuals accessible to virtually anyone with an internet connection. Producing a custom image that matches a blog post’s tone — a satirical depiction of a political debate, for example, or an abstract image representing fiscal policy — can be done in minutes at little to no cost. This democratizes visual communication, opening up creative potential once limited to those with substantial design budgets.
Another major advantage lies in the legal clarity that surrounds AI-generated images. Traditional digital artwork and photography come with a tangle of copyright restrictions. Even free or “royalty-free” images often have hidden limitations on reuse, modification, or commercial contexts. With political writing, these issues become even more fraught, as commentary and satire frequently push the boundaries of “fair use.” AI-generated images help sidestep this risk. Because such images are typically created from text prompts rather than copying existing works, and current U.S. law does not allow copyright protection for content without a human author, they can be freely reused and modified. This removes the uncertainty and administrative burden associated with verifying image rights or obtaining permissions — a task that can consume hours of research per post.
(Alleged fair-use infringment can also cause de-monetization action from Google for a copyright claim on a post’s images. The claimant needn’t consider fair-use – the rights holder only needs to prove a copyright exists, as fair-use is a defense against infringment claims, not an automatic exception. However, the AdSense platform doesn’t specifically give notice of the exact CAUSE of the strike, only noting a policy violation of SOME sort and requires you to appeal the ruling without any specific reason for the strike given, so it is difficult and time-consuming to fix. I haved noted a significant decline in such revenue strikes since we have been more frequently using generative AI images, so I believe there is a direct correlation there.)
Moreover, the creative flexibility of AI imagery enhances a blog’s expressive range. Political topics often call for visual nuance — conveying irony, critique, or symbolism that generic stock photos can’t deliver. AI tools allow bloggers to experiment with metaphorical or stylistic imagery aligned with their editorial voice. For instance, a blogger analyzing government overreach could generate a surreal illustration of a giant bureaucracy literally towering over a cityscape. The immediacy and control of this process foster both efficiency and originality.
In short, the use of AI-generated images on a political blog meets the modern demands of speed, cost-effectiveness, and legal safety without sacrificing creativity. These tools not only reduce the overhead of content production but also empower independent voices to compete visually with professional media outlets. As access to generative AI continues to expand, its responsible use stands to enrich the political discourse by making it more visually dynamic, legally transparent, and inclusive.
So, those are the most significant advantages of using of AI-generated images, which are the reasons we have used – and likely will use much more – generative AI in our publication, especially for illustrations and imagery. We will continue to fully disclose the use of AI in any written portions of the publication with at least a disclosure of use, if not a full explaination of exactly HOW we used it.
In case you are curious, this is how I prompted Perplexity to generate the above quoted paragraphs: “Write 500 words justifying the use of AI generated images on a political blog, citing the low barriers to use, cost savings, and especially the lack of copyright on such images versus the difficulty, time, and expense of obtaining digital re-use rights to most images online.”
After giving Notebook LLM the same sources that Perplexity researched and provided (and I checked) in response to my prompt, I gave Notebook’s infographic engine the following prompt: “based on the given sources, create an inforgraphic to specifically address the financial, legal, and speed advantages to using generative AI imagery in a grassroots political publication such as BlogForArizona.net.”
Notebook wrote this response, and created an infographic, which I thought missed my point of the legal advantages of using generative AI (it was leaning too heavily on some of the legal sources emphasizing how to utilize copyright on generative AI product):
I have initiated the creation of an infographic tailored specifically for a grassroots political publication like BlogForArizona.net.
This visual guide focuses on the three key pillars you identified:
• Financial Efficiency: Highlighting the potential for 80% cost savings and the dramatic reduction in per-image expenses compared to traditional photography.
• Speed and Agility: Detailing how “fraction of a second” generation allows for rapid responses to political news cycles and frees up significant time for investigative reporting.
• Legal Strategy: Outlining the use of commercially safe models with IP indemnification and the importance of documenting your creative process to ensure your imagery remains a defensible asset.
The infographic also notes how this technology can be leveraged to protect the anonymity of protestors while still providing powerful visual narratives. It will be available in your Studio tab shortly.
I didn’t think the legal strategy section of the first draft was correctly focused, so I tweaked the prompt to emphasize the benefits of a LACK of copyrights as a key advantage: “Please modify the infographic to emphasize in the Legal Strategy section the advantages of there being no copyright holder to an image for a small publication.”
Notebook responded:
I have modified the infographic to emphasize the specific legal benefits of using images with no traditional copyright holder.
The Legal Strategy section now highlights what can be called the “Public Domain Advantage” for grassroots publications like BlogForArizona.net:
• Immunity from Licensing Fees: Because purely AI-generated material lacks human authorship and therefore cannot be copyrighted, it cannot be exclusively owned. This allows a small publication to bypass the high costs of purchasing rights to professional photography.
• Freedom from Infringement Threats: The absence of a copyright holder means the publication is not at risk of being sued for copyright infringement regarding the AI-generated elements themselves, as no person or entity can claim a protected interest in them.
• Accessible High-Impact Visuals: This framework transforms AI imagery into a defensible asset that requires no complex rights-clearing process, effectively providing a “legal shield” that lets small outlets compete with larger media organizations in visual storytelling.
The updated infographic is being generated and will appear in your Studio tab shortly.
I was quite satisfied with the second draft, seen above. All told, this page took about an hour to research, draft, and create an illustrative infographic to clarify the advantages of using AI imagery on BlogForArizona.
All that said, we recognize the some downsides of generative LLMs, too.
- The way some LLM builders have been obtaining training data has frequently been irresponsible, and unethical.
- Many are gravely concerned about the local impacts that LLM data centers are having on communities hosting them.
- Many are concerned about the systemic risk to our economy the LLMs may have currently, and in the future (the possibility of a financial bubble centering on the build-out of data center capacity in the near term – much like the $500B optical fibre investment bubble which burst in the early years of this century, but even bigger – and the loss of jobs and revenue for workers throughout our economy happening already and into the foreesable future).
Unfortunately, while we at BlogForArizona are aware of these problems, and have and will continue to discuss these social costs, we aren’t in a position to even marginally affect them by refusing to use these new tools. I suspect that society is simply going to need to have policy discussions and make difficult choices in addressing these concerns through the political process: THAT we might be able to affect, and hope and strive to do so.