The tech word, the academic world, the business world and a few other worlds are having a moment over ChatGPT and other Artificial Intelligence innovations. For those not in the know, ChatGPT is an Artificial Intelligence product created by a company called OpenAI that allows users to input a query or a request. It’s trained on the bazillion words that exist on the Internet. And it can generate output that can be anything from an essay to a poem, computer code, a piece of digital art, typed text that can turn into script, something resembling an answer to a search query, to a play. Below is play I generated with the simple request: “write a short play about shepherds.” (click on the image below to enlarge it.)
Some are thrilled. (I’m sure the sheep in the story are not.) Some are concerned.
It’s illuminating quite a paradox. Educational institutions see benefits and also as a potential tool for students to cheat when it comes to writing assignments. Artists see this as potentially crowding them out of work and entire new ways for new artists to express themselves. In the journalism world we’ve already seen CNET start to use this form of AI to generate content (and get criticized for not disclosing it.) Buzzfeed says it’s going to be using it in some capacities to “personalize” some of its content.
I’m not quite sure about the PR spin on that one. Sounds like it might have been generated by a ChatGPT PR request instead of a thinking human. But then again, most PR speak is formulaic anyway.
It’s not just text either. There are now tools to generate AI digital art, music, computer code and who knows what else.
And yes, there are accuracy problems. And context problems. And plagarism problems. And…problems.
But AI is all the rage and it’s stoking rage from those that see this as a harbinger of doom vs those who see it as the future. There are some who think (and hope) this might mean the beginning of the end for Google’s search dominance. Microsoft has made a huge investment in ChatGPT prompting some to envision a more intelligent Clippy dancing across our desktops. The Chicago Tribune is wondering if ChatGPT can replace restaurant critics. Apparently Real Estate Agents are seeing dollar signs.
Maureen Dowd of the New York Times took things out for a spin and wordsmithed the description AI: Actually Insipid Until It’s Actively Insidious. (or at least her headline writer did.) That was after doing an interview with with an AI Shakespeare. Zounds!
A friend of mine, Sumocat, is generating fake Mastodon posts using a few of the tools and some iOS shortcuts to create fake Mastodon posts. (He labels them as such.)
So there’s fun and games and then there’s maybe not so fun and games.
The labor issue is a legitimate concern in the same way we’ve seen technology innovations create labor issues throughout history. (Anyone still selling printing presses?) Unfortunately no artificial or non-artificial intelligence has yet to figure out how to get companies to put some distance between announcements of large layoffs and new technology investments. Timing issues aside, the cold cruel economics of innovation always surface when something cool comes along.
I can’t and won’t judge whether or not this is a good or bad thing at this point. One never can tell how the day is going to turn out when it first dawns. But it’s a thing. And it’s a paradox.
Let me tell you a short story about my field, the theatre. Years ago electronic, synthesized and digital music technology advanced to the point where producers could replace entire orchestras with a digital track of the score. Instead of a conductor or music director all you needed to run the show was a sound engineer to push the go button and voila, the theatre filled with the booming sound of a full orchestra. Heck, you could even digitally enhance the singing during particularly strenuous dance numbers.
The change didn’t happen overnight. First orchestras started reducing their numbers as digital instrumentation meant that fewer players needed to be hired. Purists howled. Bean counters cheered. I don’t think any of that is news to anybody. But here’s the story part:
I directed a tracked musical for a dear friend and colleague. We were standing in the back of theatre watching the audience enjoy the show as it neared its final moments. He turned to me and asked what I thought of the music quality. I answered that I thought it was decent but not great and it didn’t allow the show to breathe properly. (The tracks he’d rented weren’t the best and this was in the early days of this trend.) He said that he didn’t think anyone would notice. And then he delivered the punchline: “But I hate it that I need to do this.”
The sad reality boils down to economics. It’s so much cheaper for a producer to rent tracks than to hire musicians to rehearse and play the score live. Licensing agencies are making a decent piece of change from that business.
Back in the day to engage an orchestra or a pit band or just a piano player, producers would rent the score from the licensing house. Hard copy scores would be distributed via mail or delivery service. Those scores would get used by musicians. By used I mean they also got marked up with notes, pages were dog-eared and worse. And then when the show concluded those markings had to be erased and folded pages straightened before sending them back to the licensing house.
That evolved into electronic distribution of scores, which simply transferred the cost of re-producing those scripts from the licensing agency to the producer.
The point I’m sure you’ve grasped behind this story is that innovation induces change all along any chain of production. Lives and professions are affected. Some lose out, others gain.
It seems almost inhumane to replace humans with the output of machines. But we’re well practiced at it and the almighty dollar eventually erases our memories of the tensions or dulls the pain these moments of human ingenuity create to replace humans.
But in our age of misinformation, when trusting what we see, hear, and read requires developing new skill sets because humans who have always been adept at creating misinformation have figured out how to really profit by doing so, the beginnings of using Artificial Intelligence to create “news” we see, hear, and read eerily feels like it has the potential to be too far a step over a line that maybe we shouldn’t cross.
Which ultimately is a paradox. AI is computer science. The essence of which is simply bits that are on and bits that are off. 1’s and 0’s. True. False. One would think that distinction reduced to that binary response could be helpful in sorting out fact from fiction.
It’s not. It’s not simply because it takes humans to create it. And it’s not because those humans are training these AI robots against all of the accuracies and inaccuracies, truths, lies, facts and fiction humans have created along the way.
It’s an endless paradox.
And here’s what ChatGPT has to say about that: (click to expand)
Life on the Wicked Stage: Act 3 