Blog 2: Ethical use of AI

As a socially conscious software developer, I’ve had a really hard time squaring away how I feel about generative AI models and using them. Opinions range from outright bans of AI models to the source of future human prosperity. My initial assessment during the start of this crazy was they the AI models will cap out in their ability, and that using them was kinda shitty, but as I’ve had to content with the strong encouragement to use them in the software space, I’ve had to carve out a more nuanced take.

Addressing some quick arguments

There are a few things I’m not going to get into in this article, and some grounds I’m going to set. If any of my assertions in this section don’t sit right with you dear reader, this may not be an article to read in full for you, and you may prefer to read my most recent tinkerings post.

Generative AI Is inherently unethical (for the most part)

If you are someone who would make the claim that the scanning of copywrited works, both official and social media, makes the existence of these models inherently unethical due to the fact that they should not have been able to exist, than I agree with you. I think this section will aggravate people of any extreme on this subject, but I believe that generative AI models have an inherently unethical source, and that I’m likely going to be using them anyway in some form or another. Now, this is easy for me to say; 90% of my GenAI use will be for work, and software information and tutorials have been freely posted and copied from one another for a long time. Open source, after all, tends to have a lot of people ripping answers from people with more time than them to work on more niche problems.

I believe that how you contend with this fact is up to you. If you think it shouldn’t be touched due to this, fair. However, if you choose to ignore this or make excuses, it would make someone like me doubt your integrity. However, the car appears to be out of the bag, and with the mass cooperate adoption of this technology, especially in tech, it seems I will need to pick up using it as a skill. Unfortunately.

We have many things we haft to contend with that are beyond our reach, the early colonial occupations of the America, Canada and Australia, the poor labor practices involved in almost all clothing manufacturing or almost any item you can buy, and I think learning how to contend with these issues is an important part of being socially conscious and frankly, being an adult. Some people will bend over backwards to avoid being part in perpetuating any of these systems, some will not have the resources available to them to avoid it. I admire those who do their best to avoid perpetuating these things, and I’m working to be better at this myself.

GenAI is going to fuel the Climate Crisis

The electricity use of GenAI is really not something I care to contend with as a reason not to use it. We certainly need to be making better efforts towards climate change as an inevitable problem that my generation must solve. However, I believe that this is not a problem with AI. Power usage will go up with it, and with our current energy infrastructure that will lead to more carbon emissions. Despite that, if our energy grid was clean this increase in power would not matter in terms of climate and I would prefer that those concerned with this put more effort into pushing for good energy policy, and this concern will go away.

We are all going to run out of water

I will admit ahead of time I’m not the most educated on this in relation to the data centers, but this is an opinion piece, and I will address my thoughts here. I know that golf courses take up more water than the new data centers, and I figure that if this is an issue for you, you likely have strong opinions on that. I just hope cities are able to plan well around this. I don’t think the impact will be that big, but I thought AI use would drop by now so what do I know .-.

The Meat and Potatoes

With all of that out of the way, I’m going to propose some things that in my estimations, ignoring the previous arguments are ethical, or at least more ethical uses of AI tooling.

Spite the Corporate overlords

Use of products like chatgpt and copilot etc without paying I think avoids a lot of the ethical pitfalls about how use could perpetuate these companies to continue to steal people’s copywrited works. The GenAI money pit will eventually run out, and those companies will need to make money. I use chatgpt for some things here and there, mostly idea generation, and I do not and will not pay for it. If the free tier is removed, I will stop using it. Besides that, I have use of Claude and Cursor at work, which again, I use sparingly and don’t pay for myself. The tokens were paid for on my behalf without asking me first. So yeah, if you’ve got issues with it but are curious to use it, avoid engaging in any monetary producing behavior; I say that instead of paying for it as a lot of these companies are talking about rolling out ads, and avoiding those is valuable.

I’m just asking questions maaaaaan

The primary value I’ve derived from genAI tools is asking questions in human readable formats when I don’t know enough about a subject or technology to know the keywords to search, or, when search results generally don’t tend to be that useful. I recently used it for another tinkering article to learn about how firmware works for a lot of plug in devices for windows for making a custom device, and it lead me to the HID format, which I had never heard of before.

AI Art

This is likely the most contentious part of this article knowing it exists on the internet, and I’m just going to say, AI art isn’t real art, if you generate images and call yourself an artist you are likely a fraud. I will however say, that AI generated images can likely be part of art; think generating a whole bunch of images and then forming them into a collage. For me, I generate the occasional joke image, use it for DND handouts, and may use it for temporary assets for game development. In line with being ethical, I will not pay for these generations, and I do not give feedback to the model to make it better or worse.

As a personal note, if someone uses a lot of GenAI images in their LinkedIn or their public professional work, it does make me think less of them (I see you tech influencers on LinkedIn); generally because it’s a very lazy way to add intertie, and shows a lack of social awareness on this topic. I thought about doing that for these articles as a sort of fun toungue in cheek thing, but I just never felt good about it. If you make money from generated assets that are untouched by you, I also think that is a show of poor character. I’ve seen 1500+ dnd map packs on Etsy that where clearly generated being sold for 10$, and if AI assets were left in a game after prototyping I would think less of the studio and publisher. Using these models to churn out garbage into the internet to make money is a character flaw.

Generating Ideas

This is the most personal use of use of GenAI for me and one that I think will suck the most to get criticism about. I use chatgpt like a more advanced idea generator for my dnd campaigns at this point. I use it to make traps, add handouts, give me boss fight ideas and oneshot ideas, and generally how I used to use fantasynamegenerators.com. I feel most at odds with this because I assume so much hard work of one shots from small creators that were copywrited got eaten into the machine that is LLMs. It’s always a useful spring board to spark some creativity on my part, and it can even generate some decent stat blocks (that I tend to clean up because none of them are perfect).

It’s for a personal game, I don’t pull anything word for word, and I don’t financially contribute to these models or give them feedback. This is the best compromise I feel I can give, wasting chatgpt server time on someone that will never pay them.

Open source is for everyone

The thing I feel least bad about using GenAI models for, is code. There is such a culture of give for free, use for free in software, and I think a company consuming that code that was given freely or answers given on stackoverflow are free game for scraping with a bot and using however you’d like. There’s some rumors of them breaking licenses, and that should be addressed, but that does also fall under the first section of this post. The attitude online amongst software enthusiasts and devs alike seems to be that the more we share, the better everyone’s lives can be.

Lastly, misinformation

As general advice, don’t take anything from genAI models as literally true. This is an important place to practice critical thinking with these things. I don’t trust code 100% that gets generated, don’t trust fully it’s advice in interpersonal situations, and generally take it to have a pretty flawed perspective, and it’s up to you the user of these things to extract the value from the responses to your prompts. Please use critical thinking with these things.

Conclusion

There is no ethical use of GenAI models as they were not made ethically to begin with. In our modern climate however, using generative AI models seems to be becoming more and more of a necessity, and hashing out your own personal feelings of where the line should be drawn in the sand helps show you your own ignorance and helps us all have a better conversation about these things than what typically happens on the internet. I hope my article has at the very least given you something to chew on, some food for thought if you will. If you have particularly strong disagreements with me that are not in my “Addressing some quick arguments” section, then feel free to leave a comment or email me. I am always trying to learn more on these things.