I agree wholeheartedly and appreciate you writing about AI. I’d go further with your first bullet point, “Do not outsource to AI anything that is core to your craft.” Beyond “core to your craft,” don’t use it at all in daily life except as necessary or required for your employment. Don’t treat it as a personal assistant to write emails, plan meals, create a travel itinerary, make a flyer, write an Instagram post, or any of the other things I hear my peers say they ask AI to do. That’s how the slippery slope starts and you stop thinking and creating on your own.
Thanks, Sarah. I tend to agree. The slippery slop is a real concern (and one I addressed at length on a recent episode of Cal Newport's podcast). I think what makes this so challenging is that AI is kind of like a Microsoft Excel spreadsheet in some regards, and nobody would ever say "Don't use an Excel spreadsheet to do your taxes!" But an Excel spreadsheet doesn't then creep into your life and start offering to script difficult conversations for you, and to remove resistance globally. In my own life, I use AI as an advanced version of Google search. Instead of just typing a search term into Google, I'll now sometimes phrase my search as a question, see the AI response (which Google automatically gives you) and then go to links to check for accuracy and learn more. But yeah, I wouldn't use it for much else, and perhaps I'm already on the slippery slope. (But I don't think so). I also think this is why it's so good to do things like lift weights or run or improv — real things in the real world that come against gravity that AI will never replace and that remind you how satisfying it is to overcome resistance and struggle meaningfully toward a goal.
The slippery slope is also something I've been wrestling with. I have completely forsworn AI for writing, but use it for a few other areas that kinda sorta fall into the category of "advanced Google search." For example, my rotor was rubbing on the disc brakes of my bike. I probably could have gone on Youtube and figured out how to diagnose and fix the issue, but Gemini was able to point me in the right direction immediately (well, not immediately because I have an app that imposes a 5 minute delay between when I click on Gemini and when it opens on my phone). Similarly, I sometimes use LLMs to help me use up remaining ingredients in my fridge. I've struggled to be the type of person who doesn't waste odds and ends, and have gotten better in that respect due to AI.
But I know that in both of those examples, there is probably a better way to achieve the same outcomes. For the bike, there is a local bike co-op where people will teach you to fix issues like the one I had. So I could have gotten out into my community, met people and had the skill transmitted person to person. This would have required me to spend more time on this going to/ from the co-op, as well as doing the work within its limited open hours. For the cooking example, I could put some intention and effort into becoming a better, more intuitive cook. None of the recommendations LLMs provide me are astounding. So I imagine if I spent a little extra effort, I probably could figure out how to stretch my larder and grow as a cook in the process. But at 6:30pm on a Wednesday when I'm already a little cooked myself from a long day that is easier said than done.
I think this is where I really distinguish between things that are core to who I am and want to be, and, well, other stuff, and where my time is best spent. A non-AI example: for the first 10-12 years of our relationship, my husband did all the work on our cars. He had the skills and we were broke 20-somethings: him doing the work was worth the time and saved us money. But eventually, as we grew professionally and work took up more time/energy (and gave us more money), we started taking the cars to the shop. Working on them was no longer worth his limited time (and in a very literal way - his hourly pay was more than what we'd spend at the shop).
As a high level administrator at a university, there are certain tasks that just aren't worth my time. Developing the ideas and content for a presentation? Yes, worth it (and core to my professional identity). Making the slides on the PPT deck pretty? Nope, so I use AI for that. Identifying robust and valid analytical question/questions for a project? Yes, core. Doing the analysis itself and coming up with a nice way to visualize the data? Nope.
I share a lot of Brad's concerns, and certainly all of them in creative spaces. And I appreciate the slippery slope challenges (although, philosophically I have qualms with slippery slope arguments as such). But I've seen the power of AI tools for so much of the daily, backend work of developing, implementing, and evaluating complex programs in large scale settings. Need to give faculty robust benchmarking data on other schools' curriculum? AI can pull that information for me in a tiny fraction of the time.
And to complete the thought that I hit post on too quickly: That gives me more time to make meaning of what it finds, identify areas we should align and areas where we can differentiate, and spend precious time talking to individual faculty about their ideas, concerns, etc. That is, what is core to my profession (and frankly, what they pay me for!).
I appreciate reading your words here, and I do mean, *your* words. It matters.
I just finished a book by an author and personal friend I’ve long admired, and the book was unmistakably written/revised/edited by AI, which diminished my opinion of my friend tremendously. I absolutely believe the core material was true to his experiences and the lessons he pulled from it (nonfiction), but the essence of it, his words, his arguments, was not him. (Although it was clearly modeled on his voice and style.)
My opinion of his work is diminished which makes me sad. I’ve tried to press into what bothers me so much, but I haven’t wrestled the “why” down. (Perhaps I should ask Claude? lol) But I’m deeply troubled that someone I know to be a reflective person, with a long history of thinking and expressing his thoughts, would lean into AI to this extent.
The point you made about real music being made with care, I think, gets to my sorrow. Thank you.
Thank you for writing this piece, Brad, and saying what so many feel, believe, and need to hear. Since my time as a college athlete and coach, and growing up as a deeply sensitive person, coming home to the body and all that is human is core to what helps us grow as people. I think it is so ironic that as neuroscience, medicine, and culture continue to allow more of the body into culture as a key component of intelligence and health (and not separate from our minds), this force of AI/techno-optimists continues to push an agenda that keeps the body as a uselessly slow machine, not worthy of our attention, and not part of the magic of moving forward. Thank you for your work and your voice!
Struggle is where the meaning comes from' is the line I'd keep. A finished product you didn't wrestle with is just output. The hard part is deciding what's core to your craft. For me it's the thinking and the voice. Everything else is fair game for help.
I agree wholeheartedly and appreciate you writing about AI. I’d go further with your first bullet point, “Do not outsource to AI anything that is core to your craft.” Beyond “core to your craft,” don’t use it at all in daily life except as necessary or required for your employment. Don’t treat it as a personal assistant to write emails, plan meals, create a travel itinerary, make a flyer, write an Instagram post, or any of the other things I hear my peers say they ask AI to do. That’s how the slippery slope starts and you stop thinking and creating on your own.
Thanks, Sarah. I tend to agree. The slippery slop is a real concern (and one I addressed at length on a recent episode of Cal Newport's podcast). I think what makes this so challenging is that AI is kind of like a Microsoft Excel spreadsheet in some regards, and nobody would ever say "Don't use an Excel spreadsheet to do your taxes!" But an Excel spreadsheet doesn't then creep into your life and start offering to script difficult conversations for you, and to remove resistance globally. In my own life, I use AI as an advanced version of Google search. Instead of just typing a search term into Google, I'll now sometimes phrase my search as a question, see the AI response (which Google automatically gives you) and then go to links to check for accuracy and learn more. But yeah, I wouldn't use it for much else, and perhaps I'm already on the slippery slope. (But I don't think so). I also think this is why it's so good to do things like lift weights or run or improv — real things in the real world that come against gravity that AI will never replace and that remind you how satisfying it is to overcome resistance and struggle meaningfully toward a goal.
The slippery slope is also something I've been wrestling with. I have completely forsworn AI for writing, but use it for a few other areas that kinda sorta fall into the category of "advanced Google search." For example, my rotor was rubbing on the disc brakes of my bike. I probably could have gone on Youtube and figured out how to diagnose and fix the issue, but Gemini was able to point me in the right direction immediately (well, not immediately because I have an app that imposes a 5 minute delay between when I click on Gemini and when it opens on my phone). Similarly, I sometimes use LLMs to help me use up remaining ingredients in my fridge. I've struggled to be the type of person who doesn't waste odds and ends, and have gotten better in that respect due to AI.
But I know that in both of those examples, there is probably a better way to achieve the same outcomes. For the bike, there is a local bike co-op where people will teach you to fix issues like the one I had. So I could have gotten out into my community, met people and had the skill transmitted person to person. This would have required me to spend more time on this going to/ from the co-op, as well as doing the work within its limited open hours. For the cooking example, I could put some intention and effort into becoming a better, more intuitive cook. None of the recommendations LLMs provide me are astounding. So I imagine if I spent a little extra effort, I probably could figure out how to stretch my larder and grow as a cook in the process. But at 6:30pm on a Wednesday when I'm already a little cooked myself from a long day that is easier said than done.
I think this is where I really distinguish between things that are core to who I am and want to be, and, well, other stuff, and where my time is best spent. A non-AI example: for the first 10-12 years of our relationship, my husband did all the work on our cars. He had the skills and we were broke 20-somethings: him doing the work was worth the time and saved us money. But eventually, as we grew professionally and work took up more time/energy (and gave us more money), we started taking the cars to the shop. Working on them was no longer worth his limited time (and in a very literal way - his hourly pay was more than what we'd spend at the shop).
As a high level administrator at a university, there are certain tasks that just aren't worth my time. Developing the ideas and content for a presentation? Yes, worth it (and core to my professional identity). Making the slides on the PPT deck pretty? Nope, so I use AI for that. Identifying robust and valid analytical question/questions for a project? Yes, core. Doing the analysis itself and coming up with a nice way to visualize the data? Nope.
I share a lot of Brad's concerns, and certainly all of them in creative spaces. And I appreciate the slippery slope challenges (although, philosophically I have qualms with slippery slope arguments as such). But I've seen the power of AI tools for so much of the daily, backend work of developing, implementing, and evaluating complex programs in large scale settings. Need to give faculty robust benchmarking data on other schools' curriculum? AI can pull that information for me in a tiny fraction of the time.
And to complete the thought that I hit post on too quickly: That gives me more time to make meaning of what it finds, identify areas we should align and areas where we can differentiate, and spend precious time talking to individual faculty about their ideas, concerns, etc. That is, what is core to my profession (and frankly, what they pay me for!).
Yes, especially your last point—that’s partly why I started guitar lessons at the ripe age of 57!
I appreciate reading your words here, and I do mean, *your* words. It matters.
I just finished a book by an author and personal friend I’ve long admired, and the book was unmistakably written/revised/edited by AI, which diminished my opinion of my friend tremendously. I absolutely believe the core material was true to his experiences and the lessons he pulled from it (nonfiction), but the essence of it, his words, his arguments, was not him. (Although it was clearly modeled on his voice and style.)
My opinion of his work is diminished which makes me sad. I’ve tried to press into what bothers me so much, but I haven’t wrestled the “why” down. (Perhaps I should ask Claude? lol) But I’m deeply troubled that someone I know to be a reflective person, with a long history of thinking and expressing his thoughts, would lean into AI to this extent.
The point you made about real music being made with care, I think, gets to my sorrow. Thank you.
Thank you for writing this piece, Brad, and saying what so many feel, believe, and need to hear. Since my time as a college athlete and coach, and growing up as a deeply sensitive person, coming home to the body and all that is human is core to what helps us grow as people. I think it is so ironic that as neuroscience, medicine, and culture continue to allow more of the body into culture as a key component of intelligence and health (and not separate from our minds), this force of AI/techno-optimists continues to push an agenda that keeps the body as a uselessly slow machine, not worthy of our attention, and not part of the magic of moving forward. Thank you for your work and your voice!
Struggle is where the meaning comes from' is the line I'd keep. A finished product you didn't wrestle with is just output. The hard part is deciding what's core to your craft. For me it's the thinking and the voice. Everything else is fair game for help.