What Should We Be Outsourcing to AI?
Ask yourself, "What would I make AI do if I only had one year left to live?"
I’ve been thinking a lot about AI. I think about its role in my life and the lives of my fellow humans. I think about how we should integrate it in a meaningful and useful way. Like all technologies that came before it, AI has the potential to be both wonderful and terribly awful. There is both extreme pushback against it, often out of fear and uncertainty. And there is also leaning into it to a point of failing to consider the ethical implications and major downsides of it.
So how do you start to think about how to use AI in your own life?

While many experts have a lot more to say on this topic than me, I’d like to take a new angle to help the average user think about how to integrate AI into their lives. My hope is that it opens our minds to possibilities while warning against the potential downfalls.
Death and AI – What’s the thread here?
To do this, I’d like to use a version of the death thought exercise as a good place to start.
Imagine you have just been told you have one year left to live. The only catch is that, for the most part, you must carry on with your daily life. Now, imagine you are told AI can help you with as many of life’s tasks as you can think up (and that current forms of AI can actually assist with).
What do you pick?
I’ve seen several people commenting on how we need AI to be doing our dishes and laundry, not creating art for us. I think if most of us imagine ourselves with limited time on the earth, we might feel the same way. In your final days, do you really want to outsource all your writing, reflection, and art to AI? Or do you want to find a way to use AI to free up your time so that you can spend more of it doing things you love, creating things you love, and being with people you love?
I’ve seen AI used to do a number of amazing things, and I’ve also seen it shortcut things that either humans are better at or benefit from doing. For instance, I’d love to see AI help organize my notes, cull the internet for data to summarize for me, or help draft outlines or checklists. But I don’t necessarily think AI should or could be thinking of the next cutting edge research question on how to improve engagement in end-of-life care planning (my area of research).
Similarly, I think AI can help inform doctors about how to elicit patients’ preferences for end-of-life care, but I’m not convinced it should be having actual end-of-life care conversations.
What about research?
To test my theory, I sat down with ChatGPT to ask about advance care planning. Advance care planning is the process of planning for medical care you receive at the end of life. It includes both having conversations about the care you want (with family, loved ones, and medical team members) and filling out formal documents outlining what you want. It’s my core area of research. We know there are major benefits to completing advance care planning, so the biggest mystery my field has grappled with is – how the heck do you motivate people to do it?
ChatGPT did a phenomenal job at curating a very comprehensive and easy to understand set of information about advance care planning for me. I could tell it was doing a great job of pulling and organizing data from my field. In fact, dare I say, it did a better job than me the expert at explaining a lot of the terms in accurate but easy to understand ways. This is something that might take me several hours if not days to do, but it was done in seconds.
But when I asked ChatGPT how to help motivate me or help me plan, it fell drastically short. It just kept iterating on various versions of, “just do it.” Or “follow these steps.” This isn’t Nike, so that slogan isn’t really helpful in the context of overcoming my fears, avoidance, confusion, or myriad of other barriers to completing the task of advance care planning. In other words, it had very little to say about how to motivate humans very specifically in this case.
So, the organizing of information was absolutely within the domain of AI, but the execution of how to utilize that information and/or ask meaningful research questions was something where at least to date, AI falls greatly short of a human expert.
This is no surprise to any of us, but a useful example of what about research can be executed by a machine and what still needs to be done by a human.
What about art?
If you ask an AI art image creator, like DALL-E, to create a painting of a dog, it probably won’t come up with much that is interesting. What makes imeages created by DALL-E or other AI art tools good, if anything, is the knowledge, ingenuity, and creativity of the prompts that human users bring to the tool. It’s in co-creating that the art becomes any good at all. I’ve played around with it, and it can come up with some seriously underwhelming and often really off-putting art without good guidance and direction. Also, hands…why can’t AI get hands right??
Certainly, the field of AI is developing rapidly and there are many tools outside the OpenAI network to play around with. Some of it very fun, and some of it very sophisticated given proper prompts and tools.
Where I gain pause is when we start to use AI to replace core human capabilities that make us happy as humans. For instance, creating art actually increases the levels of dopamine in our brains. Dopamine is a neurotransmitter associated with pleasure and reward. Where the research on the effects of art on the brain start to get really interesting is that the process of creating any form of art (despite medium and talent level) can reduce stress, improve mood, and increase self-confidence. Even if you are afraid your art will be terrible, research shows you still benefit from the process of making it.
Christianne Strang, who is a professor of neuroscience at the University of Alabama Birmingham, notes that “creativity in and of itself is important for remaining healthy, remaining connected to yourself and connected to the world.” So, it’s the process of creating art that yields a lot of benefit for us as humans.
So how does AI play into art making? The concern for which I think more research needs to be done is the immediacy with which you can create art using AI tools. Like much of our “you can get it right now” culture emerging with the advance of technologies, AI tools apply the same approach to making art. You can iterate through hundreds of paintings in less than a minute if you want, without the work and process involved in the more standard routes of creating art.
It turns out, much of the benefit of creating art is in the very process of creation. For instance, one study in the Journal of the American Art Therapy Association found that 45 minutes of creating art in a studio lowered cortisol levels (i.e., the “stress” hormone). Creating art also helps you focus deeply and get into what is referred to the “flow” state, characterized by theta brain waves and deep levels of focus. Repeatedly and rapidly hitting command prompts into an AI tool likely don’t yield that same benefit.
Going back to my original thought exercise – imagine if you had a year left to live, what kind of art would you create? Would you be motivated with your little time left to sit in front of your computer and practice developing multiple prompts to have AI create the art for you? Or would you sit in a studio, go into nature, hit a pottery wheel. Which would bring more pleasure? Which would matter more to your kids or loved ones if you left it behind?
Although AI is relatively new to the scene, I can say not a single dying person I have worked with or known has said they wanted to sit down and do much of anything glued to a screen.
So, if you need a quick image for a post or materials to throw into a document, I think AI image generators are the perfect tool. But if you want to benefit from the act of creating art, then the traditional route serves you best.
What about forming connections?
I have seen AI used to generate rapid-fire content for posts, feeds, comments on posts, and connection reach outs. Of course, there are major advantages to short cutting some of this work, but you must be careful not to start replacing the art of forming meaningful and genuine connections with others with trying to fit in as much volume as you can. These are short and shallow connections with limited meaning rather than deeper, more meaningful connections (where humans thrive). Humans weren’t really designed for rapid-fire and widespread connection with other humans at a bot-like rate, and this trend is something to watch.
As noted before in prior newsletters, good social relationships are the single best predictor of happiness. Of course, you might be getting these in your day-to-day life while you build out rapid fire connections via social media platforms. However, be cautious how you build out your networks and the degree to which they are socially or professionally rewarding.
There is now a plethora of research to indicate that the degree to which social media platforms can yield benefits is the degree to which you make meaningful social connections on them. So, if you connect with your loved ones or friends on Instagram or TikTok, or perhaps even if you are just entertained, you may derive some benefit. The degree to which those things are absent is the degree to which you see diminishing returns from using these platforms.
Once again, ask who and how you’d build out your network if you knew you had limited time left to live. You’d probably seek out groups related to your own interests or needs or people with whom you felt it was meaningful to engage. You might pull back on the rapid-fire platform building that our culture seems to praise.
Of course, if you are building a brand or a business, you do need to grow your network. But there are ways to do that more meaningfully, slowly, and carefully that might yield greater benefits in terms of happiness. So, by all means, use AI to help you brainstorm what to post about or topics of conversation, but be wary of outsourcing all your connecting and relationship-building. You might be missing out on the greatest benefits of these social network platforms.
Bottom line
Things will change rapidly as AI evolves, but I think the take-away here is that AI can be used to shortcut a lot of stuff you already don’t like to do. What a great way to free yourself up to do more of what you love. Summarizing a person’s resume for me, coming up with interview questions, making checklists or to do lists – all of these fall within the domain of things I would absolutely love to have AI do for me.
But anything where the process of it is pleasurable and rewarding – creating art, forming relationships and connections, interacting with others – AI not only does a worse job than most humans but outsourcing it steals from us a unique opportunity for joy.
Reflection questions/exercises
Time to put AI to the test this week and have some fun with it! If I had to summarize the barometer for when to use AI or not, it’s this:
Are you having fun with AI? If so, GOOD!
Is there something you don’t really like to do, and AI could do for you? GREAT!
This week, reflect on these questions to see how you could use a little more AI and be a little more human this week:
Think about one place or way in which AI could assist you this week and try to give yourself some time back.
o Could it write a grocery list for you? (Hint: feed it all the recipes you want to do in a week and ask it to create a grocery list)
o Could it come up with a to do list?
o Could it create a chore chart or exercise plan for you?
o Could it help you decide what are the best plants to plant in your climate?
o Could it help you plan a vacation?
Think about something this week you want to savor as being completely human?
o Can you draw or sketch a piece of nature for 30-45 minutes? (bonus tip: you don’t have to be good at it at all!)
o Can you cook something new?
o Can you garden?
o Can you create something? (Writing, woodwork, painting, drawing, etc.)
How can you balance meaningful connections with short cutting other tasks (e.g., summarizing your resume)?
May you enjoy how superior we are at being human this week!
For fun, here’s a really great list of 35 ways people are using AI from the New York Times to get you started.
*Note: I am aware that talking about AI is one of the more complex and nuanced discussions of our time, and this is a very simple summary of some of my thoughts on it. I have many thoughts on the ethics, possibilities, limitations, etc. But for now, this will work for a short newsletter. Always feel free to share your thoughts or resources!
I enjoyed your thinking and pondering what else AI can/ought/should do.
Personally, my enjoyment lately has been writing the prompts and the code (and having AI assist as well) to integrate my "second brain" apps on my computer (note-taking apps, databases, and other file repositories). Tweaking the prompt (or sequence of prompts) to get the most optimal output takes time. AI doesn't know what you expect of it, so the more specificity and parameters you give it, it can tailor your request. I've also noticed Claude does some things better than ChatGPT and Gemini better than Claude, etc. Lots to discover! Me: I wonder if I can do this… and there goes another few hours.
Your title is a great way to frame this "what do we really want AI doing" in our personal lives question.