I just finished watching 60 Minutes. Tonight’s episode was on Artificial Intelligence and how it is about to be unleashed on the world. Last night I had a conversation with my son about AI, and it was news to me that there were already programs available that will write you a story, or a procedure manual, or paint you a lovely picture. That conversation and 60 Minutes both sort of frighten me. I had no idea AI had come that far and was poised to take over.
Earlier in the week I heard a news story that said a large group of scientists, including Elon Musk, had signed an open letter asking AI developers to stop for six months, but that news item just seemed like background noise I could safely ignore. The unsettling part of the 60 Minutes story, for me at least, was that AI can tell lies, and frequently does. It showed an AI bot being asked a question, which it answered quickly. The answer included references to a half dozen or so books a person could go to for more info, except not one of the books actually exists! The AI just made up some legit-sounding shit like the human receiving the answer was to be trifled with. The story went on to say that AI lies were very common. So what happens to the truth? What if AI starts making up conspiracy theories or expounding well known BS like election fraud? It still won’t be true, but it will have more of an air of legitimacy because it will appear to be researched and cross referenced. So instead of trickles of nonsense coming out of the keyboard warrior’s basement lair, it will be coming from the research source itself. Take the volatile political environment in the US today. It really only got out of hand when truth was questioned (fake news, alternate truths) and fringe ideas got a national pulpit to preach from. What if a few million people who already believe the government should be overthrown get some fresh fertilizer from their AI bots, who suggest they amass and coordinate? And the humans who created the AI are only on the sidelines already, by the look of it. The AI is teaching itself to do stuff, then turning around and coding and programming the next computer in line, and no human is in between making sure they are virtuous and honest. And if any company stops for six months like the scientists want, then the company down the road they are competing with will get a six month head start and surpass them.
What will become of writers? NYC Midnight, the place that runs the short story competition I went in a few times, have a new rule in their contests that states if they detect your story was written by AI you will be disqualified. On 60 Minutes, the AI they used wrote very convincingly and I think what it wrote would be hard to sniff out as fraud. Worse of course, is the fact some poor bastard will write, edit, re-write, rinse, repeat for weeks to make a good story and years to write a decent book. But his AI counterpart can crank out a compelling story in a few seconds, and could literally write thousands of books in the same time. If AI would be a good sport and sign itself as the author of whatever it writes, then no harm done I suppose, but the technology has already show itself to be a serial bullshitter. And some guy could get his bot to write him a million dollar idea for Hollywood just for the asking, so why get all sweaty trying to be honest?
Maybe the biggest irony will be that the programmers who created AI will be the first job category eliminated by their invention. It will be the real life equivalent of Frankenstein’s monster coming to life and choking his creator.
Frankenstein by Mary Shelley is still relevant today. It’s amazing to me that book was written by a girl of 18 and was published in 1818, without any help from her tech.