A good piece from The Atlantic discussing recent AI capabilities that are beginning to point out what many have been talking about for some time: this is a different dynamic.
Historically technology or automation has replaced lower skill jobs, or at least jobs that didn’t require higher education or a lot of “creativity.” The creative arts were often thought of as being safe from “machines,” as the lore was that humans were special when it came to things like writing, painting, making music. But over time, that specialness has been slowly eroded. In music, tools like auto-tune, digital audio workstations, and text-to-music have changed the landscape, but still have been viewed as accessories with the musician being the creator.
Do capabilities like ChatGPT and Dall-E represent a sea change, or just more evolution of tools for creatives to use? There are arguments on both sides, but frankly I’m seeing a lot of head in the sand (hubris?) on the creative side of things. I asked ChatGPT about the difference between SN1 and SN2 reactions and the answer was as good or better than any of my students over years of teaching organic chemistry. While talking with a group of university art students, we challenged Dall-E to a rather odd prompt and one of the five images generated could probably have passed muster.
Technology has a history of changing the nature of work, and a history of evolving the role of the human. But these changes pre-digital (e.g. the printing press, assembly lines, radio/tv) had slower adoption, giving society a bit more time to adapt and adjust training, education, and job markets. Digital changed all of that, and that pace of change seems to be exponential. A year or two ago, plenty of hallway conversations about AI, with many saying the job loss claims were over-rated, and we were still years if not decades away from “real” AI. While ChatGPT and Dall-E are certainly not sentient or the be-all of AI, hard for anyone to deny how quickly these models have improved, and how steep their trajectory likely is.
There is an up-side to this. AI promises to improve much of the human conditions. From better energy efficiency to better healthcare to better education – these are all in-play. But it will also disrupt every industry it touches, and it will touch them all. And given our track record, highly likely that the touch will have serious ethical and equity issues. All the more reason to be working through those angles now. These capabilities are an opportunity to revisit what it means to be human, what it means to be machine, and what the relationship is between the two.