I’m an AI optimist.
I even write a newsletter about AI.
You’re reading it right now. How meta!
I believe in AI’s potential to change our lives for the better by making us more productive, creative, informed, and other rosy things.
At the same time, it’s hard to turn a blind eye to all the decidedly less rosy things: AI content that’s borderline plagiarism, questions about how LLMs are trained, and more.
We’ve seen a growing number of legal challenges and court cases with AI companies being sued by artists, journalists, and actors.
And—as any other powerful technology—AI also helps shitty people be shitty at scale, whether it’s deepfakes, spam, or scams.
In a world where all of the above is true, can using AI at all be ethical?
I think so.
You might not.
But what we think is largely irrelevant.
Judging by the way the AI train is chugging along, we’re slated to see more AI in our lives, not less. It’ll become ubiquitous in every digital product you use.
In the medium-to-long term, I’m sure we’ll find ways to capitalize on the promise of AI while limiting its downsides, whether it’s through targeted regulation, clearer guidelines, fair compensation for creatives, or other means.
Until that utopia arrives, we’ll just have to live in this ethical gray zone and navigate it to the best of our abilities.
Here are some principles I try to follow with AI to keep things as “white hat” as I can.
Maybe you’ll find them useful, too.
General principles
Don’t break the law.1 Using AI doesn’t make you less accountable. AI scams are still scams. AI-enabled copyright infringement is still copyright infringement. Et cetera.
If you’re using AI content for profit, consider disclosing how AI is used.
If you’re using AI for shits and giggles, you can afford to be more lax.2
Be ready to take personal responsibility when sharing AI-generated content with a wider audience.
Reflect on and adjust these principles as AI and the landscape around it evolve.
LLMs
Don’t release AI-generated content into the world without human review and insight. There’s more than enough low-effort fluff out there.
Don’t use LLMs to spin content wholesale without adding value.3
Limit how much sensitive and personal data you share with LLMs about other people. Take steps to minimize exposure if you do need to share such data.
Double-check any factual information from LLMs before sharing it. Be conscious of LLM “hallucinations.”
Similarly, be aware of any inherent biases in LLMs.
Don’t use LLMs for shady stuff. I mean, come on.
Creative AI (images, music, video, etc.)
Limit the use of specific artists (especially living) in your prompts.4
Label AI content if shared in forums where it might otherwise mislead people.
Credit original creators if using their work as a direct reference for AI.
Don’t monetize copyright-infringing works made by AI.
Don’t generate harmful or illegal stuff.
The choice is yours
If you read my newsletter, chances are you’re also an AI optimist and excited about what it can do.
I feel it’s up to us to showcase the responsible ways of using AI and serve as examples of the better future we’d all like to see.
You have little say in how others choose to use this promising tech.
But you have control over your own actions.
I think that’s a good place to start.
Over to you…
How do you feel about using existing AI tools? Is it a free-for-all or do we all bear responsibility for making sure it makes the world around us better instead of shittier?
Do you have different opinions on what constitutes ethical AI use? What principles do you personally follow? I’d love to hear your thoughts!
Leave a comment or shoot me an email at whytryai@substack.com.
This should go without saying, but here we are.
I often use known characters and movie references in my Midjourney prompt series. I treat it as “fair use” as I’m not profiting from it. But I’d be curious to hear a counter-opinion.
I used a living artist, Nychos, in the very first Midjourney prompt I ever shared. Since then, I’ve intentionally focused on generic prompt modifiers that don’t rely on artists’ names.
Is there an ethical way to use a dictionary? Yes to both. And both could be citing where their information comes from (words are discovered in the wild, taken and used without permission), but much easier for AI companies. So why don’t they?
This is a good start. AI ethics also goes beyond these use cases. You have to be careful about the use of deepfakes (audio and video), the automating your thinking, using private data, and respecting cultural and ethical differences between societies.