Living and Learning: A Personal Journey in Generative AI Ethics
AI has been a part of our lives for longer than many realize. Any iPhone user who appreciates the quality of their photos can thank AI for enhancements like on-device image processing. For over a decade, tools like the Photos app and Google Photos have allowed us to search for "green car" or "snowman" and more easily find the image we're looking for, powered by an AI ability to recognize and categorize images. These everyday conveniences existed well before the recent buzz around generative AI. In fact, the roots of AI trace back at least to the 1960s[1][2], with Lisp introducing declarative paradigms - focusing on "what" to do rather than "how" to do it. Perhaps because of this long history and the many ways it has been applied, the term "AI" remains loaded and often unhelpful, as it oversimplifies a complex and evolving field.
I’ve worked with AI, and on AI products, for about a third of my career, in non-generative contexts - and have in a way become numb to the term because of the overhyped and imprecise use of the term "AI" over the past five decades. When generative AI arrived, it felt like an inevitable next step, albeit a predictable and somewhat uninspiring one. Tools like GitHub Copilot were already handling mundane, boilerplate code before ChatGPT hit the scene. As a result, I started using generative AI tools almost reflexively, without giving much thought to their broader implications.
The example I want to focus on is the previous "artwork" on this blog, including my avatar and the blog banner. It was all pixel art generated with ChatGPT/DALL-E 3 because I didn’t want to spend the time creating it myself. I’ve done pixel art in the past and still dabble occasionally, but as an amateur game developer, I find it particularly time-consuming - I’m a better coder than artist. In fact, I’ve previously commissioned pixel artists to create sprite sheets for my games. Since I don’t blog often and my site doesn’t get much traffic, it felt like an obvious choice to generate the artwork instead.
Over time, as I spent time in game development and other artistic communities on Mastodon and Bluesky, I began to reflect on the real impact of my choices. I started to see how decisions like using AI-generated art could affect others and what those decisions reflected about my own values. These realizations didn’t come all at once, but they gradually forced me to confront uncomfortable questions about the consequences of prioritizing convenience over humanity and craft.
Let’s start with what these choices do to others. Many talented artists and photographers create digital assets for modest fees, often relying on this work to support themselves. These individuals, many of whom dedicate their lives to mastering their craft, deserve opportunities to thrive. When products like ChatGPT and DALL-E hit the market with feverish hype and fanfare, the disruption to their livelihoods can be significant. Industries like game development, digital illustration, and stock photography have surely already felt the effects, as AI tools make it easier to bypass human creators. The implications for these communities could be, or perhaps already are, profound - affecting their ability to sustain the standard of living that we all strive for.
Not only that, but what does it reveal about the person prompting AI to make art? What does it say about their understanding of "art" and its value? Is art reduced to something functional, worth accepting even when it contains obvious flaws that a human artist would avoid? What does it say about someone who chooses AI-generated images over creating something themselves or supporting a skilled artist - someone who could bring passion and expertise to their work? These choices reveal our priorities and how we value creativity, effort, and the human condition.
I’m not sure how these questions resonate with others, but they hit me hard. What am I saying about myself as a neighbor, citizen, and human if I believe my blog post needs artwork, yet I’m unwilling to create it myself or support an artisan? The answer was clear: I was being lazy - too lazy to invest the time in creating a pixel art image on my own. Additionally, I was inconsistent about the value of my blog. If I rely on a hollow shell of what once was a deeply human concept, "art," what does that imply about the substance of my intellectual contributions? It was a sobering realization. But the realization that perhaps hit the hardest: My actions revealed an indifference to the people most vulnerable to the AI revolution, prioritizing convenience over concern for human dignity.
This post is my mea culpa. Upon some reflection and realizing the implications of my choices, I removed the AI-generated banners from my blog and replaced my avatar with a real picture of myself until I get my hands on a sweet new pixel art portrait from a professional pixel artist whom I supported[3]. I hope this is a strong first step towards living by the values I want to uphold.
The undervaluing and displacement of artisans; the casual redefining of a deeply human endeavor as foundational as art; these are only a couple aspects of the ethical questions we should be asking ourselves, and I am just beginning my own journey in exploring them. I still use AI tools for tasks like code completion, debugging, and design ideation, but the broader implications are hard to ignore. There’s the environmental cost of running these systems[4], the representational biases that generative AI often reinforces[5], and the reality that AI relies on unlicensed use of intellectual property[6]. On top of that, our government seems horribly unprepared to address the societal challenges posed by AI. Instead of engaging in meaningful discussions, the US gov't perpetually gets bogged down by populist rhetoric and performance. Consider the potential impact of self-driving trucks alone - potentially displacing 3.5 million truck driver jobs in the U.S.[7] - and the need for thoughtful policymaking becomes clear.
Life is full of complexities, and ethical questions rarely have simple answers. The choices we make - whether in creating art, supporting others, or adopting new tools - deserve thoughtful reflection. I invite you, my friend, to sit with these questions, to weigh them carefully, and to navigate this ever-changing landscape with care and discernment.
[2] https://www.redhat.com/en/command-line-heroes/season-3/talking-to-machines
[3] https://ko-fi.com/bachelor_soft/commissions
[4] https://www.wired.com/story/true-cost-generative-ai-data-centers-energy/
[5] https://www.ibm.com/think/topics/shedding-light-on-ai-bias-with-real-world-examples