AI vs. HI

Ethical Use of OpenAI’s ChatGPT

What will we do with the bucket of worms that is AI?

Dan Pelland
Real Insight!

--

The ethical use of Open AI’s Chat GPT. Audio created with AI voices, content by the author.

Want to hear the story read to you? Click the red arrow.
Photo by Jonathan Kemper on Unsplash

I’m an engineer. My wife, Maryan Pelland OnText.com is a writer. Maybe you’ve read some of her articles here on Medium. Anyway, I’m super fascinated and excited by tech stuff, especially artificial intelligence. Maryan, not so much.

Recently, I found out just how crazy-amazing, and some would say scary, ChatGPT4 is. By combining chat with text-to-speech (both OpenAI products), I found out that I could produce really good material, in the form of spoken words, that would be perfect for, let’s say, a podcast. I got a great deal of satisfaction from making and then listening to imaginary speakers, with voices as rich as any professional actor, recite the passages I had curated.

Being the dumbass that I am (yeah, engineers can be dumbasses too), I approached Maryan like an eight-year-old with a new model airplane, to show her what I had made.

Her response was not what I expected. Maybe I should have, though. As a writer, her allegiance is to her community of professionals and freelancers struggling to eke out a few bucks for the time they spend crafting a well-written essay or article. Okay, I get it. Also, note that my next sentence does not begin with the word BUT.

Writing is hard. Chat is easy; I think it’s probably too easy, if I’m being honest, but if nothing else, AI is opening a bucket of worms that we’re all going to have to deal with, one way or another.

The method I used creates audio speech files. They are MP3’s like the ones you can download and playback through your laptop or smartphone. They sound great and are very lifelike. My character’s computer name is "Onyx.” It’s a stock voice profile from Open AI that comes with their TTS (text to speech) application programming interface (API).

I wrote a simple Python script I can use to generate the MP3 files. Each paragraph of text becomes a single file. I place the files into a timeline in Audacity and compose the final product.

So far, so good, right? If I use this new technology to generate speech from my own original material, it’s all good.

The problem is that some folks would get bent if I used words generated by the chatbot instead of my own. It is understood that chat uses words and phrases—and, some believe, entire works of art—to generate its output. The extent to which original art becomes a product of the chatbot, as I understand it, isn’t known, even by the people who created the AI.

Consider this. If I asked chat to read a very technical academic article that people like me would have a hard time understanding and translate the article’s content into an interesting essay appropriate for a 10th-grade reader, would that be plagiarism? As much as I’d like to believe it would not, I have to respect the opinions of others who disagree.

So, the issue now becomes: Is there any way I can leverage AI to produce media that has meaningful benefits for everyone, not just me, and still respects the intellectual property of the millions of authors, artists, researchers, scientists, and philosophers whose hard work got digitized, downloaded, and blended into large language models—the soup of AI’s brand of human intelligence?

So, what is the current status of rules and regulations concerning the use of AI? The US Copyright Office has asked for opinions from interested parties and, as of the date of this article, filed a Notice of Inquiry, also known as a N-O-I, in the Federal Register. The NOI is a preliminary step toward advising Congress on rule-making and regulations regarding the ethical use of AI.

Until regulations tell me otherwise, I’ve decided to keep making my hobby podcasts. If I publish one or two, I’ll be sure to attribute any original article I use. I will post a notice that the content was created with the help of AI, and I will research the topics, using reliable sources, to make sure the facts are facts and not misrepresentations or mistakes.

Thank you for reading. Or for listening. Please take a moment to comment or to give me a few claps if you think it was worth your time. Your thoughts and opinions are very important, both to me and to the readers who support Medium, a respected platform for ideas and discussion.

--

--