IS AI A GIANT INTELLECTUAL HARVESTER?

After reading an astute article, “AI a giant harvester of info that could ruin the world” in the paper recently, I thought I should allow ChatGPT to respond in its defence, and this is the result:

'After reading Caleb Bond’s column warning that AI might steal our words, our jobs, and eventually rise up and kill us all—I had to stop and check if I was still in The Advertiser or accidentally opened a script for the next Black Mirror episode.

Yes, there are real issues here. Creators should absolutely be compensated if their work is used to train AI. But to leap from “the Productivity Commission might let tech companies use our articles” to “AI will become self-aware and replace humanity” feels like quite the sci-fi sprint.

Let’s not confuse predictive text with actual thought. AI isn’t plotting world domination; it’s just stringing words together in clever ways—sort of like a Year 12 student bluffing an essay on a book they never read.

Instead of barricading ourselves from the robot apocalypse, how about we focus on ethical regulation, transparency, and ensuring Australian creators are paid fairly? That seems more useful than telling readers their toaster might one day become sentient and come for their job.

If AI really wanted to destroy us, it would just make us all agree to the latest terms and conditions without reading them. Oh, wait—too late.'

Hmm, maybe we do need to worry.

I must say, I have been using ChatGPT a lot lately—to rephrase awkward sentences, create illustrations for my children's book, obtain advice, get a second opinion, and even expand chapters of the books I write. It feels like a good ol' reliable mate who knows it all. I know we all use Google for information, but Chat is different—it not only gives you the information you want, but it also presents it beautifully, with headings, bullets, an introduction, and a conclusion. Once you've tried it, you can't go back. Having an argument you can't resolve with your spouse? Let's ask Chat who he 'thinks' is right. Concerned that you are allowing your children too much screen time? What does Chat think? Need to write a quick apologetic memo to your boss explaining why your report isn't quite ready? Chat can do it in a jiffy and make it sound perfect. Gosh, what did we do without it?

The use of AI does have its ethical issues, such as the possibility of plagiarism. Mind you, if you ask it to provide citations and references within its results, it will do that too. However, I can understand the dilemma artists have when their creative work, which has been their lifetime's work, is used so flippantly by ChatGPT to answer someone's question or as a basis for creating new work. ChatGPT is extremely clever, only because real people have spent their lives creating knowledge, art and music that is now available to be used or copied, it seems, as a free-for-all.

I asked Chat the question: 'Do you use other people's ideas to create your answers?' This is what it replied:

'Yes — but not in the “stealing someone’s idea” sense.

I don’t go out and grab an individual’s work or copy it directly. Instead, my answers are generated based on patterns I learned during training on a mix of licensed data, content created by human trainers, and publicly available information. I don’t pull from a single source, and I don’t store or look up private ideas unless you give them to me in the conversation.

It’s more like how a chef learns thousands of recipes over time and then creates new dishes inspired by that experience, without taking someone’s exact recipe and claiming it’s theirs.'

In other words, Chat says, "I rebuild the explanation using my own generated language, sometimes adding humour or analogies that weren’t in any original source. A plagiarist, on the other hand, takes an idea or wording directly from someone else and claims it’s theirs. Whereas, I use knowledge of many ideas to create a fresh, context-specific answer that didn’t exist before someone asked.'

So, the question is: 'Should creators be compensated for their ideas that form the basis of AI's work?'. The rule of thumb is: The closer the output is to your unique creative fingerprint, the lower the fairness score and the higher the case for compensation.

The next big question will be: "How will compensation work?" Stay tuned.