Policy & Analysis


Insights
ChatGPT For You and Me
By
Dr. Sabrina Hoque
Unlike pretty much every promise ever made about AI, ChatGPT and its cousin DALL-E actually deliver on their potential. This is a good thing, since they make it possible to increase the output of skilled workers—something to which we will all benefit.
Artificial Intelligence (AI) has had many false promises over the years, with futurists like Ray Kurzweil promising for decades essentially the emergence of intelligent machines similar to the character Data of TV’s Star Trek, all to no avail. Even autonomous driving, the subject of billions of dollars of investment and frothing speculation of venture capitalists, performs an awful lot of test drives but still shows no sign of any mass—or even niche—market appeal, with no path forward.
This is why the recent emergence of ChatGPT, an AI chatbot trained to articulate detailed answers that mimic human conversation for pretty much any question you can come up with, is so impressive. It can write rap songs, poetry, play games, and complete exam answers at above the level of the average human test-taker. Combine this with the remarkable DALL-E 2, an AI from the same firm (OpenAI) that creates stunning digital images based on subject prompts (“Draw me a unicorn riding a tiger in the style of Monet”), and for the first time ever natural language AI offers true promise.
ChatGPT makes factual errors and the writing is hardly Shakespearean. But then, that describes most humans. The quality of the product is so good professors have begun to worry students will use it to cheat on their homework and essays. There is no doubt use is already rampant in schools. Many educators are on the verge of panic, fearful students recourse to cheating will sabotage their ability to learn. Employment effects, too, are on the minds of many. Economist Paul Krugman worries the affect on skilled workers in particular may be negative. The thing about a tool that really works is that we now have to consider real effects.
While it is important to think carefully about these questions, needless worry about new technology is an incredibly ingrained, highly wasteful human trait. Waxing poetic about the writing skills of students before AI is not something done by anyone whose read a handful of undergraduate essays. If students are going to use the new tool it is because they need the help. And as for concerns about knowledge workers, it is true that some may be displaced—that almost invariably happens. But for most, AI is likely to only enhance their writing, providing a rough draft or even just suggestions that can accelerate the pace of output, enhancing productivity and leaving more time for other things that the skilled are good at.
We should, then, view the arrival of useful AI as a good thing, if not worthy of celebration, then at least a commitment to taking advantage of all that it offers.