STORIES WE TELL OURSELVES

Legalities vs Ethics: We’re Still Arguing About AI-Generated Stuff

Developing laws takes forever, but what of fairness and ethical behavior?

I’ve been a professional writer for more than half a century, and I can tell you how much work writing is, how easily people steal your work, and how challenging it is to make a decent living.

I’ve written books, newspaper stories, magazine articles, web content, marketing stuff, grants, and radio scripts. I have ghostwritten about a dozen books. I’ve made a great living.

My firm stance is this: A person who cannot compose a relatable, clean, supported, engaging, useful piece of writing needs to be in some other form of expression. That person is not a writer.

No AI platform or trick will ever make them into a writer. If their work can’t be born independent of a robot stuffed with material purloined from human creators, it is not their work and is unethical.

On writing and human creativity

So, we consider the “content developer” who prompts a bot to write an article about widgets — something the particular human is not qualified to write about.

We think about the entrepreneur who sees an opportunity to scrape dry academic articles from a website and run them through a paraphrasing bot so the end product is more palatable and “written” for a 5th-grade audience.

The bot uses a database crammed with the work products of human creatives who did not give permission and were not compensated. Then the AI platform and the individual prompting the bot make money, but the writer or illustrator does not.

The human involved may vet the material or claim to. But does that person have the knowledge, the experience, or the research skills and tools to verify what the bot cranks out? Does that person understand what a primary source is when writing factual prose? What qualifies this person to sell such work and affix their name to it? Are these products plagiarized?

A court would have to determine that. To date, in the EU and the US, qualifying to register a copyright requires proof that a work was substantially made via human creativity — and writing a prompt does not suffice. The applicant must disclose to what degree the work was impacted by AI generation. Do you think applicants are being truthful about this?

On any writing platform, in any publication, in any classroom setting, it is not unlikely to find AI-generated or “paraphrased” content lurking in the shadows. Peruse Medium.

In any ten stories, at least one is not 100% human creativity. There are, for example, many stories that seem fairly well written, but it’s obvious that the writer has not much background in the topic.

One comes to mind: a biographical sketch of Amelia Erhardt by a 20-year-old writer whose first language is not English. It’s a dull but not terrible story, and the language seems polished. The “author” posted several similar pieces daily and bragged about his income.

The writer’s responses to comments on the stories are, however, written by someone with a tenuous command of English. This writer and a few others recently disappeared from Medium after a group of readers pointed to the problem. The writing was not human writing.

Ethics vs. legalities — are we seeing talent and skill or just exploitation?

For me, it isn’t as much about legalities as about ethics and understanding how these AI platforms are developed.

In 2022, as all of the big ones were being launched, anything available on the web, whether copyrighted or not, was used to force-feed the databases, like geese for fois gras.

• No authors, artists, or other creatives had an opportunity to opt-out.
• Writing was just vacuumed up and used as the basis of huge windfall profits for the AI industry.

If you listen to the Copyright Office webinars from last spring, participating creators almost immediately found the demand for their work decreased sharply, and the demand has not recovered. Small publishers went out of business because they could not compete with the avalanche of AI-generated material.

Back in the early 21st century, freelance writers could make a very decent living working for publications and websites. Then, offshore content mills sprouted everywhere, paying writers pennies per hour to churn out barely literate stuff. Established writers had to compete with hacks in a boiler room who churned out poor-quality but really cheap content for around $5 per 1000 words.

Freelance writing, for 15 years or more, became an impossible way to make a living. Writers like me had to reinvent their processes and clientele to survive. The market slowly recovered to some extent as clients learned that their reputations suffered.

Then came the winter of 2022. The world went nuts for artificial intelligence, and platforms sprang up everywhere. Now, anyone can provide any content they wish by riding the backs of every writer whose work has been dumped into databases. Everyone is now “creating” something with AI.

I call your attention to the hundreds of opportunists who are more than happy to help you make thousands of dollars a month, even if you lack experience or skill, by simply feeding prompts to AI platforms and selling the results as your own work.

Those “mentors” make a fortune charging others for the “secrets.” I expect they’re happy to share a bit of their windfall with the AI platform of their choice. Win, win. Lose (for the creative worker).

What’s the solution?

For now, there isn’t one. There are too many people eager to convince themselves that producing writing or images that are not the product of their skill or talent or intellect is fair use.

Laws are coming. The EU AI Act now requires that generative AI like ChatGPT comply with transparency requirements:

• Designing the model to prevent it from generating illegal content
• Publishing summaries of copyrighted data used for training
• Disclosing that the content was generated by AI

Most AI platforms do not comply with such regulations, especially the second one.

There are numerous lawsuits pending in the US and EU against every one of these AI platforms, from Open AI to Google to Github and Microsoft. It is worthwhile to read who is suing and why in this deep-dive article:

On Medium, and on every other publishing and content venue, this is a complex and serious issue that has made a lot of money for people exploiting a process they don’t understand.

These folks call themselves writers and illustrators while they game the entire industry. But they’ll tell you with an ingenious demeanor that they only seek to disseminate information to everyone and make the world a better place.

But what they’re providing isn’t their work product. It’s opportunism at best.