Embracing AI: My Journey with Image Generation
- Sep 21
- 6 min read
Updated: 4 days ago
By Audrey Korte
This summer, I had a costly incident with AI.
One would think I would never want to use it again.
Correct — but I am exceptionally aware of my lack of experience with AI and want to understand this stuff better.
The Lightship site was created using Wix and Canva, with some help from three web designers. They mostly aided me in getting the site secure and functional after a cybersecurity incident. It is still a work in progress, but the site is operational again.
The cybersecurity incident is why this post was delayed by days.
I am glad to be back, but my disappointment over the botched launch and security breach is substantial.
My First Encounter with AI
What I want to focus on today is my first interaction with an AI image generator. You may notice that the lead image on my landing page was made with AI.
How could a person who lost their job for an AI-related mishap use it on their website?
It’s actually a bit of a metaphor.
I decided to use it because it speaks to the issues I write about but also looks pretty groovy.
When I was looking at stock photos, I wasn’t happy with the options.
They felt too generic. There were images of hurricanes, boats, a ship's wheel, and general water photos. There were lighthouses. But nothing came close to what I wanted.
But lo and behold, there it was — an AI image generator option.
That’s new. It's been a minute since I built a blog site.
“Okay,” I figured.
“May as well see what it can do.”
I admit I don’t know a lot about AI, but I’m willing to learn.
An AI image generator is a technology powered by Artificial Intelligence that can produce images, artworks, or visual content without direct human intervention. It employs complex machine learning algorithms based on deep neural networks trained on vast datasets of existing images, according to Fortune Business Insights.
“These neural networks learn the patterns, styles, and features of the images during training,” the article says. “Once trained, a generator can take various inputs, such as text descriptions or random noise, and produce new images that align with the input.”
AI image generation started in the 1960s when pioneers began exploring computer-generated endeavors, far earlier than I would have guessed.
“Early AI art projects in the 1960s and 70s were limited because they depended heavily on algorithms and preset rules. However, the integration of networks and machine learning in the 1980s and 90s greatly opened up the possibilities for AI-generated art,” according to Global Trends in AI Image and Art Generation: A 5-Year Analysis Across Countries.
According to a PhotoGPT blog post from April 2025, the global AI market was worth $279 billion in 2024 and is expected to hit $1.8 trillion by 2030, with image generation playing a major role in creative industries.
AI-generated content now powers over 68% of the images used in marketing and social media campaigns, according to PhotoGPT.
And 34 million AI images are created daily, with users ranging from small business owners to global creators.
A couple of weeks ago, I became one of those AI image creators.
Making the Image
I started by asking for a ship in a storm at sea with a warning light on it. I asked AI to make the ship old-fashioned.
That left me with a new image that was somewhat improved. Each version was pretty different from the last. Even when there were components I wanted to keep, I couldn’t always figure out how to do that.
I specified that this was a lightship, and that seemed to help, surprisingly — the light no longer looked modern at least.
A lightship serves like a lighthouse at sea, an anchored warning to sailors in fog and dangerous waters. It’s a metaphor for my place in the news industry at the moment.
I tried again.
I added that there should be a woman on the ship in stormy waters. At one point, the woman was on the water. Next, she was underwater.
When I got her on deck, the light I wanted atop the ship instead looked like it was coming out of the lady’s head.
At one point, it appeared as if lightning bolts were coming out of her eyes.
I got that sorted, but the woman looked like she had been in a shipwreck. Too worn out and defeated.
Together, AI and I got her dressed a little nicer, in an old-fashioned manner, in red — my favorite color.
I added glasses to her face.
But she was too skinny.
Sigh.
Eventually, I got a woman who looked kind of like me, where I wanted her. I added that she needed a newspaper under one arm and a camera around her neck, in reference to the journalism industry.
The newspaper’s text did not include any English, but then I asked it to make an American newspaper. Still, the words on the page were mostly nonsensical, with the exception of “American” in the headline.
And that’s where I stopped. I thought it was a little 👅 tongue-in-cheek commentary on how the blog began, but so too, the lack of exacting results. The image is neato, but it was a process to create, and I never got exactly what I wanted.
Now, I am not trained in AI prompts. I’m learning. I don’t yet know if there’s one prompt that could’ve gotten me from point A to point X in one fell swoop.
Learning AI is a process. I’m just beginning. I don’t want to remain so ignorant about the subject and feel compelled now to understand more after avoiding it over the years.
The lack of knowledge meant I found some things out the hard way. .
According to the World Economic Forum’s Future of Jobs Report 2023, AI and automation will lead to both job displacement and job creation. The report suggests that 69 million new jobs could be created by 2027, while 83 million existing roles may be displaced, highlighting the need for workforce adaptation. The impact on creators is already being felt as people elect to use AI to create images over hiring graphic designers, for example.
I worry that AI is going to erode human oversight of content while we fast approach a more generic online world that is growing less accurate and accountable. I wonder how much it will change in our daily lives and careers over the next couple of years alone. Beyond that I can’t imagine. It’s still so new to me.
Honestly, in terms of art and graphics I prefer the human-designed. When I have the money, I will hire a graphic designer. But not having that ability currently makes image generation more attractive but it also means endless potential for devaluing the creative human-made work. I believe in employing humans.
It is easier to personalize and provides a product that is more unique and creative when you make the time and money for it. It also requires planning and finding creators you like, trust and who get your vision. Doable? Definitely. But also one more thing to consider as an experimenting blog creator with a ton on the to-do list.
Obviously I have mixed feelings about the whole thing, but I am determined to keep learning, cautiously and tentatively, because I don’t think AI is avoidable. It’s already crept into my life in ways I wish I could escape. Being unprepared did not help me avoid complications — it made assumptions and misunderstandings easy.
The Importance of Knowledge
Knowledge is power. It is important to take some time to figure it all out. Sometimes that means training. Sometimes that means playing with AI. The trick is to do it with tact and attention and to not bring it into a high-stakes situation where errors can do irreparable damage.
This was a pretty fun intro to AI image content. While I appreciate what I was able to produce with an image generator, it is a bittersweet reminder of what being uneducated, untrained, and unprepared for artificial intelligence can cost a person.
As I continue this journey, I hope to share more insights and experiences. The world of AI is vast and ever-changing, and I want to navigate it with confidence and clarity.
Comments