Assistive vs. Generative AI: Where's the Ethical Line?

What a question, right? The topic of AI is a hot button issue on several fronts:

• AI's impact on climate change and damage to the environment

• AI's impact on the working class

• AI's impact on copyright and intellectual property protection

I recently did a podcast episode on these three aspects of Artificial Intelligence and it's definitely worth the watch to get a deep dive into them, which I'll summarize below. But you should also listen (on Spotify or wherever you listen to The Neurodivergent Creative Podcast) or watch on YouTube!

The Environmental Impact of AI

How could something I am doing on my computer that takes ten minutes be impacting the environment or climate?

This is a reasonable point of confusion. It seems weird to think that one person's use of a technology has such an impact. But when you examine the scale of that tech use considering all those who are using it, on large and small scales... it becomes clearer. AI requires a lot of processing power - which means it produces heat and needs to be cooled, thus using VAST AMOUNTS of our fuel and water resources.

This article from the Harvard Business Review by Shaolei Ren and Adam Wierman explores the environmental costs of AI usage - specifically focusing on the ways those costs are inequitably dispersed around the world. Here's the highlights:

• "...the training process for a single AI model, such as a large language model, can consume thousands of megawatt hours of electricity and emit hundreds of tons of carbon. This is roughly equivalent to the annual carbon emissions of hundreds of households in America."

• "AI model training can lead to the evaporation of an astonishing amount of fresh water into the atmosphere for data center heat rejection, potentially exacerbating stress on our already limited freshwater resources."

• "...the global AI energy demand [is] projected to exponentially increase to at least 10 times the current level and exceed the annual electricity consumption of a small country like Belgium by 2026."

• "In many cases, adverse environmental impacts of AI disproportionately burden communities and regions that are particularly vulnerable to the resulting environmental harms. For instance, in 2022, Google operated its data center in Finland on 97% carbon-free energy; that number drops to 4–18% for its data centers in Asia. This highlights a significant disparity in the local consumption of fossil fuels and the creation of air pollution. Similarly, the water-consumption rate for data center heat rejection can be disproportionately higher in drought-stricken regions such as Arizona due to their hotter climates."

As a human being who cares about the environment and the survival of all species including my own, I want to do as much as I can to reduce my carbon footprint, conserve natural resources, and not be a blight on the planet. But what can one person do? Is my own avoidance of Generative AI enough to reduce the environmental impact of its use? Probably not. But that also doesn't mean I'm just going to dig in and use it because 'why not?'

We are living in unprecedented times. Asheville, North Carolina has been all but wiped off the map because of Hurricane Helene. ASHEVILLE. That is not a coastal city. That is in the mountains. Far inland. We have to stop pretending that climate change is just part of a natural process and isn't anything we could stop. Yes we could, but not with corporations and billionaires calling the shots. And not with Generative AI being baked into every piece of tech we use in our daily lives.

The Economic Impact of AI

As reported by CNN, a poll of 2,000 executives from 18 industries in nine countries - including the US, Canada, Germany, and Japan - revealed that over 40% of companies expect to employ fewer white-collar and blue-collar workers due to AI.

• "Some 46% of executives said they would redeploy employees internally if their jobs were impacted by AI. And two-thirds said they planned to recruit people skilled in AI compared with just over one-third who said they would train their existing workforce in the technology."

• "In the past year, some tech firms, including file storage service Dropbox and language-learning app Duolingo, have cited AI as a reason for making lay-offs."

• "Goldman Sachs economists said in March last year that as many as 300 million full-time jobs could be lost or diminished globally by the rise of generative AI, with white-collar workers likely to be the most at risk."

The Washington Post also reported on the billions of dollars being invested into the AI bubble and the economic impact as tech companies try to recoup their investments and secure monopolistic positions as leaders in the space that will choke out smaller startups.

• "The immense cost of training AI algorithms — which requires running mind-boggling amounts of data through warehouses of expensive and energy-hungry computer chips — means that even as companies like Microsoft, OpenAI and Google slowly begin charging for AI tools, they are still spending billions to develop and run those tools."

• "Beyond the Big Tech companies, a legion of start-ups is trying to find ways to make money with such 'generative' AI tech as image generators and chatbots. They include trying to replace customer service agents, writing advertising copy, summarizing doctors’ notes and even trying to detect deepfake AI images made by other AI tools. By replacing workers or helping employees become more productive, they hope to sell subscriptions to their AI tools."

• "In the United States, smaller AI companies have expressed worries that AI leaders such as Google and OpenAI will lobby the government to make it harder for new entrants to compete."

Generative AI is based on theft, full stop

The tough part is here! This is the area where people really dive into the gray area of Generative AI and start fighting in back alleys over it. LLMs, Language Learning Models, like ChatGPT, have to be programmed. Fed. Taught. They have to eat lots of content to understand it, and then they use what they learned to generate whatever you ask for. Whether or not it's even correct.

And they may even be training themselves further on your work that you input, or the documents saved to your Google Drive, or your expertise and presentation style captured by Zoom. CHECK YOUR TERMS AND CONDITIONS, PLEASE!!!!

My Ethical Line in the Sand

It is going to be very hard to fully divest from Generative AI considering how many daily use programs have added it as a tool. But just like I put gas in my car and I throw my to-go cups in the trash because my city doesn't recycle that kind of plastic and I buy pre-chopped veggies in plastic bags... I cannot ask myself or my clients to completely eschew a thing that is 100% categorically, provably Bad For Humanity. Generative AI is here to stay and largely unavoidable.

But in the same vein, I can carpool and run multiple errands in one trip to reduce my fuel usage. I can turn off lights and faucets when they aren't in use. I can compost my food scraps to reduce our household's carbon footprint. I can meal prep when I have the energy. I can do things to REDUCE my dependence on those Bad For Humanity things that still help people and provide accommodations.

And ACCOMMODATION is really my personal line. I work with chronically ill, mentally ill, autistic, ADHD writers. We are disabled. We do not have the same 24 hours in a day as someone of the dominant neurotype and/or without chronic illness. So yeah, I have to get okay with the fact that Generative AI can indeed be a helpful accommodation that levels the playing field.

But I am unwilling to work with someone when ChatGPT is doing the bulk of their creative work for them.

So when I work with clients and partners, these are BIG NOs:

• Using a LLM such as ChatGPT to generate content that is claimed and credited as one's own work

• Using a LLM such as ChatGPT to alter or summarize my (Caitlin's) work that I have written myself

• Using a Generative Image AI such as Midjourney to produce visuals made from the stolen work of artists who have not been compensated or even asked to use their work

And these are uses of AI that I support for my clients:

• Using a transcription AI such as Otter or Descript to transcribe audio

• Using a LLM such as ChatGPT to summarize or reorganize one's own work (though I don't think ChatGPT does a very good job of this, but that's not my line to draw)

• Using a LLM such as ChatGPT to find the word you're looking for (something mentioned in a Thread that I thought was a great example)

• Using assistive AI to double check grammar such as passive voice, punctuation, etc.