AI as intern

In the studio talking to my imaginary assistants pic.twitter.com/P7lgqVZQBN

— Austin Kleon (@austinkleon) March 22, 2023

When Kevin Kelly visited me in the studio last week while he was in town for his SXSW talk, I asked him to tell me about AI, something I am not terribly interested in. (One of the pieces of advice in his forthcoming Excellent Advice for Living is: “For a great payoff / be especially curious / about the things you are not interested in.”)

He told me that AI right now is like having a little assistant to boss around and make you some stuff so you can say, “Most of this is garbage, but I can use this part, and you’ve given me something to work with or against.”

Quoting him from elsewhere:

This first round of primitive AI agents like ChatGPT and Dalle are best thought of as universal interns. It appears that the millions of people using them for the first time this year are using these AIs to do the kinds of things they would do if they had a personal intern: write a rough draft, suggest code, summarize the research, review the talk, brainstorm ideas, make a mood board, suggest a headline, and so on. As interns, their work has to be checked and reviewed. And then made better. It is already embarrassing to release the work of the AI interns as is. You can tell, and we’ll get better at telling. Since the generative AIs have been trained on the entirety of human work — most of it mediocre — it produces “wisdom of the crowd”-like results. They may hit the mark but only because they are average.

This, so far, has been the most convincing case I’ve heard.

But then, I’ve always resisted having an assistant — in my experience, doing the “grunt work” of researching, writing a first draft, etc., is where a lot of my good discoveries are made. I want my hands on the work, because that’s how I find it.

(Though I do like bossing Siri around and telling her to remind me of stuff and I do like being a good assistant to my future self.)

More notes from our visit: