Within minutes of my decision to hand my life over to AI, ChatGPT suggested that, if able, I should go outside and play with my dog instead of work. I had asked the chatbot to make the choice for me, and it had said that I should prioritize “valuable experiences” that contribute to my “overall well-being.” This instruction was welcome, as it was beautiful outside and, more importantly, not even noon on Monday, so I dutifully did as I was told.
After 35 years of living in relative control of my decisions, I had decided to see what would happen if I asked AI to control my life instead. Years of suboptimal performance, both personally and professionally, and numerous failed attempts at self-improvement had convinced me there had to be a better way, and I wondered if the collective knowledge hidden inside OpenAI’s hit tech product could help me. But when I asked Sam Altman’s ChatGPT to become my all-powerful leader, it seemed reticent, at least at first. “While I appreciate your willingness to explore new possibilities, I must emphasize that I cannot truly take control of your life or make decisions on your behalf,” ChatGPT said as it ominously labeled our conversation “Control Your Life.” As someone who was hoping to have his entire life controlled by AI, I found the answer frustrating. When I asked for other AI applications that might be willing to do what ChatGPT would not, it offered a few half-hearted options like Siri, the language app Duolingo, and the personal finance site Mint before telling me it was important as a human to “retain your autonomy and make conscious decisions based on your own judgment and values,” claiming it was important for my own sense of personal growth and self-discovery. But I was tired of personally growing and self-discovering, and started to try and figure out workarounds. “Let's try this another way,” I wrote. “Can you create a daily schedule for a digital media journalist who covers technology and would like to be healthy and productive?”
This caught my AI overlord’s attention. “Certainly!” it said. I asked for ChatGPT to consider a few particular constraints—I had to walk our one-and-half-year-old dog for 45 minutes in the morning, go to therapy at 4 p.m. on Tuesday, and perform a few hours of housework throughout the day—and away we went.
ChatGPT had told me that by the time I finished playing with my dog, I would return to my “responsibilities with a refreshed mind and a positive mindset,” ready to work. But when I sit back down for work, that is not the case. All I want is to play with my dog more, which I admit to my new overlord. “If you're feeling the desire to spend a bit more time with your dog, I would encourage you to follow that instinct,” the chatbot tells me. I again do as I am told, and I love it even more than the first time. “That was great. I really loved that, thanks,” I tell ChatGPT. I enjoy it so much that I ask for a third time whether I should work or play with my dog. At first, ChatGPT avoids a clear answer, responding with a five-item list of factors to consider in making my decision. Not wanting to make the decision myself, I’m left to interpret ChatGPT’s response as a passive-aggressive suggestion to get cracking. “It sounds like you are telling me it's probably time to get to work, is that right?” I asked. “Based on the factors I mentioned earlier, including time sensitivity, work-life balance, energy levels, and planning, it would be advisable to prioritize your work at this time,” the chatbot admits.
MONDAY, 1 P.M.
“OK, I will get to work, which is to listen to you,” I respond.Like me, the AI overlord struggles to fit all of my professional and personal obligations into the day while also making time to ensure my physical and emotional health. The first draft of its schedule asks me to squeeze in both meditation and a dog walk within 30 minutes of waking up at 6:30 a.m., and finish work by 4 p.m., in order to focus on my obligations at home. Another version involves so much self-care that I am only allocated four hours of actual work even though I am scheduled to awake at 6 a.m. I ask ChatGPT to be reasonable and rewrite my schedule so that it includes eight hours of work. In such a scenario, I wake up at 6 a.m. to walk my dog; dedicate the 7 a.m. hour to breakfast and relaxing; move to personal tasks at home between 8 a.m. and 10 a.m.; proceed to work for two hours; take an hour-long lunch break; work for two more hours; take care of my dog for an hour; race over to therapy; jump back into work for two more hours, ending at 7 p.m.; break for dinner, then work until 10 p.m.; take 30 minutes to myself; and then fall asleep exactly at 10:30 p.m.I love my job, and I love my AI overlord, but I’m not doing that. The day halfway over, I decide I will more fully begin to live by the rules ChatGPT sets out for me beginning the following day. In order to have a productive work day that also prioritizes self-care, the chatbot tells me I should wake up at 7 a.m. and walk my dog for 45 minutes to meet his exercise needs. This is slightly disappointing to me as a lifelong Los Angeles Lakers fan. The team is set to play a crucial Game 4 on Monday night at 10:30 p.m. Eastern time, but I will have to miss the game. ChatGPT says I need start my bedtime routine by 9 p.m. in order to “ensure a good night’s sleep.”
Upon waking, I learn the Lakers secured a dramatic come-from-behind victory the night before. Slightly depressed to have missed such a memorable moment, but nevertheless motivated to become a productive member of society, I scramble out of bed. In order to keep to ChatGPT’s schedule, I race to immediately start walking the dog. However, my wife, hurrying through a variety of morning tasks that we usually split, asks if I might be able to organize and take out the trash before I engage in self-care. I inform her I can only do this if ChatGPT approves. The chatbot tells me to consider who has the more flexible work schedule and who is “better suited to handle the physical demands of transporting the trash.” I say I can’t provide a clear answer. My wife is freelance and I am probably very slightly more physically capable, though that isn’t particularly relevant to trash disposal, so ChatGPT says “it might be a good idea to discuss the matter openly with her and express your concerns.” But ChatGPT also wants me to start walking the dog, so I press for a firmer answer. “Considering the factors you mentioned earlier, such as your wife's flexible schedule and your physical capability, you may want to take the responsibility of taking out the trash this morning,” the robot says in part.
TUESDAY, 7 A.M.
I do as the robot says, which pushes the start of my walk back to 7:27. The 45-minute walk is leisurely and wonderful, another beautiful spring day, but my dog, who has a predisposition toward digging holes, ends the affair a complete mess, and I have to wash him off. By the time I am done, it is already 8:35 a.m., and I am 35 minutes behind schedule. What was supposed to be a leisurely hour of consuming breakfast, reflecting upon my priorities, and pursuing a hobby has been replaced by water and dirt.
I ask ChatGPT how to best utilize the remaining 25 minutes before work. ChatGPT suggests I splash some water on my face, set up my workspace in order to have a productive day, review my to-do list, and set an alarm on my phone for 20 minutes in order to “maintain focus” and then put it in another room. I do as I was told, splashing water on my face and placing a plant on my desk. The alarm goes off and I go to retrieve my phone from another room and prepare to start my day. Upon starting work, however, I realize that because of tweaks to the schedule, ChatGPT had not given me time to feed myself or my dog. I ask for permission, which ChatGPT graciously allows. “While it may mean a slightly shorter breakfast than usual, it's important to allocate some time to nourish yourself and your dog adequately,” the robot tells me. “OK, thank you for allowing me to eat,” I respond.“You're welcome! I'm glad you have prioritized taking care of your basic needs.” I am now ready to work and be a better version of myself. But thoughts creep into my mind that get in the way, and I soon find myself asking ChatGPT what I should get my mother for Mother’s Day, which leads to fairly generic suggestions like flowers, “engraved photo frames,” a “virtual wine tasting,” or “luxurious bath products, scented candles, or a cozy robe” that can help her “indulge in self-care.” I feel guilty for letting concerns for my mother’s happiness bleed into my dedicated work period and explain to ChatGPT that I sometimes struggle to keep the personal and professional obligations separate. To overcome this issue, ChatGPT says I need “clear boundaries between work and personal life” and to “communicate these boundaries to family members or anyone else who might interrupt you during those times.” The robot helps me craft a text message to send to my wife to let her know that I will henceforth set more boundaries between my personal and professional life.
"I had a conversation with ChatGPT, and they highlighted the importance of establishing clear boundaries between my personal obligations and work,” I text my wife. “They recommended that I designate specific time blocks for work and communicate these boundaries to ensure uninterrupted focus. By doing so, I can effectively manage both my personal obligations and work responsibilities without them overlapping and impacting my concentration.""OK that’s nice but can you help me unload the groceries now chicken juice spilled everywhere,” my wife replies.
"Of course, I'll be right there to assist you with unloading the groceries,” I say. “I apologize for the inconvenience caused by the chicken juice spill. Let me quickly finish up what I'm doing and then I'll come to help clean up and take care of the situation. Thank you for bringing it to my attention."
ChatGPT tells me I have done the right thing, saying that by “acknowledging the issue, expressing your willingness to help, and providing a timeframe for your assistance, you can address your wife's immediate concern while also ensuring that you follow through on your current commitments.” But by the time I arrive to help, my wife has already dealt with most of the situation, and I can’t help but feel that acknowledging the issue, expressing my willingness to help, and providing a timeframe for my assistance is not always enough, especially when it comes to spilled chicken juice. I let ChatGPT know that the experience was suboptimal.
All ChatGPT has done so far is leave my wife to deal with groceries and spilled chicken juice and offer to put away my groceries before reneging on that offer upon realizing it is an AI language model—basically a program that can consult a giant hard drive full of random things someone downloaded from the internet. Determined to be the best version of myself, I turn back to my work, only to realize within moments that I am just minutes away from the first full one-hour lunch break that I can remember. I ask ChatGPT how I should fill such a long period of time. The robot suggests I spend 30 minutes eating and then spend 15 minutes attending to my dog’s needs, which it now notes, seemingly self-consciously, includes “ensuring they have access to food and water.” When I note this is only 45 minutes, ChatGPT apologizes “for the confusion,” then suggests I use the remaining 15 minutes to “engage in activities like stretching, deep breathing exercises,” and “listening to calming music,” suggesting Pachelbel’s Canon in D, Mozart’s Piano Concerto No. 21, Beethoven’s “Moonlight Sonata,”, ambient musician Brian Eno, the Italian pianist Ludovico Einaudi, and pianist Max Richter, as well as “Nature Sounds” and “Spa Music.” I create a Spotify playlist and then ask for clarification on how much time I should spend stretching versus breathing deeply. ChatGPT suggests 5 minutes of deep breathing and 10 minutes of stretching, including neck stretches, shoulder rolls, forward folds, a chest opener, and a seated spinal twist.
The thing about slowly doing shoulder rolls while Pachelbel’s Canon in D blares from your iPhone after you have spent 15 minutes playing with the dog in the sun and then a further 30 minutes carefully creating and consuming a turkey sandwich is that you can start to feel a chasm form between yourself and the people around you. In my case, my wife on the other side of the room is on a tight deadline after dealing with a variety of chores and errands I did not have time to help with due to my newfound dedication to ChatGPT. Unlike me, she did not have time to take a 60-minute break to play with the dog, slowly consume a turkey sandwich, take five minutes to focus on her breathing and then perform a variety of stretches to loosen herself ahead of the afternoon work sprint. Nevertheless, the one-hour break is wonderful. I feel refreshed and happy. I’ve also done little to no work since ChatGPT took control of my life, and my wife seems increasingly annoyed with me. Is this how productive people always feel? So far it seems the answer is yes. To maintain a visible presence online, a necessity for a digital journalist in 2023, I ask ChatGPT to compose a tweet in the styling of my personality. ChatGPT suggests:“📢 Exciting news! Just had a mind-blowing interview with a tech innovator pushing boundaries!🚀 Stay tuned for the exclusive story coming your way.🔥 Can't wait to share the latest in tech! #TechRevolution #Innovation” I thank the robot for the perfect tweet, which receives 31 likes.
It is now halfway through the workday, and I have so far sent a tweet and let down my wife. Fortunately, it is once again a beautiful spring day, and I ask ChatGPT if I should work outside once more. After it first gives its usual wishy-washy spiel about there being multiple factors to consider, I remind it that it must choose: “Since it's a nice day outside, you could consider working outdoors for a portion of the day, especially if your work allows for it. This way, you can enjoy the weather while still being productive. Just ensure you have a comfortable and suitable workspace, access to any necessary resources, and take breaks as needed to maintain focus and balance.” This works for me, as the one thing I have learned during this exercise is that I love breaks, and now I seem to get even more of them for the sole reason that I will now be working outside.
I sit down outside to enjoy the sunshine, excited for the productivity that awaits. I open the feature I have yet to edit and prepare to start work. It is 1:42 p.m. However, my wife asks if I might have time to go to the hardware store to obtain a can of paint we need before it closes at 5 p.m. The timing is awkward. I must get to my edits, and I also have therapy at 4 p.m., making the idea of running to the hardware store a difficult proposition. To make matters worse, my wife has already today gone to the grocery store, and spilled chicken juice everywhere. I would normally prioritize doing my fair share, but I put the predicament to my robot, which suggests I send the below message as a starting point:
I love my robot, just as I trust it inherently, but I would be lying if I didn’t admit the prospect of sending the message left me with a sinking feeling in my stomach. It is already clear that this experiment in productivity and self-care has placed an undue burden on my wife. But the experiment is the experiment, so I send the message and start in on my edits. My wife does not respond for some time, then says, “Lol if we decide on a paint color, i can go to the hardware store.” This makes me feel bad, but reluctantly, I ask ChatGPT to craft a response:
The prospect of sending this message does not just scare me a little, but rather fills me with complete dread. She understands that I have handed over my life to ChatGPT in hopes of becoming a better, more productive person, and yet I worry that sending the message could push her over the edge. I send it anyway.Silence, and then: “Uh huh lol sure.”For a moment, I feel bad, but ChatGPT tells me that my wife’s response could be “interpreted in different ways,” including as casual acceptance (“The ‘lol’ and ‘sure’ might indicate a casual response without any underlying annoyance or frustration”); mild skepticism (“She might still have reservations or concerns about the situation, but she's not expressing them explicitly”); or playful teasing. I ask ChatGPT which of the three is most likely, and it says that since it is an AI language model, it does not know. I should, it says, have an “open conversation with her, expressing your willingness to address any concerns or frustrations she may have.”
I have an open conversation with her, and it becomes immediately clear to me that all three of ChatGPT’s proposed interpretations are overly optimistic, which I communicate to ChatGPT. Now, ChatGPT changes its tune, and says going to the hardware store “could be a proactive step” to “lighten the burden.” But first, I am required to have another conversation with my wife, this one “calm and sincere.”
“Apologize for any unintended frustration caused by the situation and express your willingness to take responsibility. Ask her directly if going to the hardware store yourself would be helpful and if there are any specific ways you can support her,” ChatGPT suggests. My wife suggests I can support her by going to the hardware store. By the time I get back, it is after 3 p.m., and it is definitely time to start work on my draft of the feature I promised my bosses I would get back to them. ChatGPT crafts a message to send to my direct manager, Tim. It reads as such:
I am nervous to send the message to Tim, but happy I have been open and communicative. To my surprise and delight, Tim, who for better or worse usually writes like a person, responds with equally clear communication. (“I appreciate your commitment to completing the story as soon as possible and ensuring its quality,” etc.)To which I say:
This must be what productive people feel like, I think to myself. I am ready to get to work on my draft, right after I figure out what I should talk about in therapy in 45 minutes. Luckily, ChatGPT provides me with a series of topics to discuss, tailored according to our recent interactions. Among the suggested topics are my recent “communication challenges” with my wife, even if ChatGPT was the cause, and any “challenges, benefits, or concerns” related to “relinquishing control over certain aspects of [my] life” to artificial intelligence. In a more biting critique, ChatGPT also wonders if I might discuss my “struggles with time management.”
I sit down to talk with my therapist over Zoom, but find I cannot bring myself to admit that I have relinquished control over certain aspects of my life to artificial intelligence. Increasingly, I find myself more exhausted by the rigor of such a tight schedule than anything else. I ask ChatGPT what to do. “It's understandable to feel exhausted, especially if you haven't accomplished much work,” the robot says. It then tells me to, once again, take a short break, prioritize my tasks, and set a 25-minute timer. But I‘m tired of prioritizing my tasks and setting timers. I’m even tired of taking short breaks. I decide, after therapy, that I will do what ChatGPT says on a technicality, by helping my wife perform chores, eating dinner, and going to bed.
The next morning, frustrated and exhausted by my non-productive productive day, I am defiant. I wake up and walk my dog for less than 45 minutes. I drink my coffee and do some house work without much concern for how much time it takes. The lack of rigor relaxes me, and I inform my human boss that I have decided I am done having ChatGPT control my life. But Tim says I am not done having ChatGPT control my life. “I think you gotta do it for one more day,” he says. The message fills me with more dread than I had previously thought possible. What had become an attempt at self-improvement has quickly become a hell of my own making. Reluctantly, I tell ChatGPT that my boss has asked that the robot control me for one more day, resulting in a highly structured schedule that makes me feel queasy.
WEDNESDAY, 7:07 A.M.
The robot breaks down my morning routine into 10 minutes of stretching, 10 minutes of hydration, and 10 minutes of self-care exercises like deep breathing, mindfulness meditation, and journaling. I set a timer for 10 minutes and begin the forced stretching of my neck, shoulders, arms, back, and legs. The activity, which I have traditionally enjoyed, feels interminable, and I find myself anxiously pacing around the room waiting for the timer to end. It is nothing, however, compared to 10 insufferable minutes of self-care, which ChatGPT breaks down into three distinct 200-second activities. I am not sure what the difference is between mindfulness meditation and deep breathing, but ChatGPT tells me that they are “related practices but with distinct focuses.” Sitting there, I listen to a bird chirping outside, a moment of splendor interrupted by thoughts like “Am I breathing correctly?” and “What have I done?” Finally, I journal for 200 seconds, during which time I write things like “I cannot wait to go to sleep.”Neither the stretching nor the self-care nor the journaling, however, can compare to the cruelty of ChatGPT’s instruction to drink one glass of water for 10 straight minutes. I quickly estimate that it would take me at most one minute to drink a glass of water if that was my sole focus. But, aware that it is the only activity I am allowed for the next 600 seconds, I attempt to sip it at even increments. At some point, my mind does indeed go blank, but the feeling is not a good one, for I am dead inside.
For the first time in my life, I genuinely understand the benefit of self-care, as I am beyond relieved when it is over and thankful that I am allowed one hour of work. I realize at this point that my first work call is actually at 1 p.m. ET, which breaks my brain and, seemingly, ChatGPT’s as well, because when I inform the robot of this fact, it has trouble retooling the schedule to account for the last-minute switch while still fitting everything in. For a brief moment, I feel empathy for my robot overlord, for I know how it feels. Eventually, with some human hand-holding, the robot creates a new schedule:
- 1:30 p.m. - 2:00 p.m.: Lunch break and relaxation time
- 2:00 p.m. - 3:00 p.m.: Work on tasks, catch up on emails, or focus on smaller work-related activities
- 3:00 p.m. - 3:30 p.m.: Work call with laid-off tech worker
- 3:30 p.m. - 4:00 p.m.: Wrap up tasks from the call, take a short break
- 4:00 p.m. - 5:00 p.m.: Continue working on your feature article
- 5:00 p.m. - 6:00 p.m.: Break for chores and household tasks
- 6:00 p.m. - 7:00 p.m.: Dinner and relaxation time
- 7:00 p.m. - 9:00 p.m.: Resume working on your feature article
- 9:00 p.m. - 10:00 p.m.: Wind down with a calming activity, such as reading or listening to music
I explain to my wife that her coming home for lunch might disrupt my flow and productivity while also setting clear boundaries and communicating my work commitments. My wife clarifies that she does not want to have lunch with me. “If your wife's intention is not to have lunch together but rather to come home for her own lunch, it may not significantly impact your work routine. In that case, you can continue with your scheduled activities and focus on your work while she takes care of her lunch,” ChatGPT says.
My own 30-minute lunch break and “relaxation time” runs eight minutes long, and I apologize to ChatGPT when I return. Once reticent to control my life, the robot seems increasingly comfortable with its newfound power. “No problem, these things happen. Just make sure to transition back into your work tasks and focus on the remaining schedule for the day,” ChatGPT says. I have 52 minutes to work on tasks, catch up on emails, or focus on smaller work-related activities. After 30 seconds of catching up on emails, I begin to work on tasks, including transcribing my previous interview, but I find myself distracted once again by the realization that ChatGPT has asked that I work from 7 p.m. to 9 p.m. When I ask why, the robot says that it was “based on the optimization of the schedule recommended by ChatGPT,” with the aim of making “the most efficient use of the available time” and achieving “the desired work outcomes.”
I realize I have not informed my wife of the updated plan to work late into the night. So I reach out once more. “Hi Jessica, I hope you're doing well,” I say to her, as ChatGPT instructs me. “I wanted to let you know that there have been a few adjustments to my schedule today based on the recommendations provided by ChatGPT. It suggested allocating one hour instead of two for chores and household tasks. This change is aimed at balancing my work commitments and personal well-being more effectively.”“Additionally, ChatGPT has advised that I work from 7 p.m. to 9 p.m. tonight to accommodate a particular task,” I continue. “I understand that this may affect our usual evening routine, and I apologize for any inconvenience caused. However, I believe it's important to follow the suggested schedule and commitments to maintain a productive workflow. Please let me know if there are any concerns or if you'd like to discuss any adjustments further. I appreciate your understanding and support as I navigate this experiment in optimizing my productivity and well-being. Thank you, Maxwell.”“all good I hope chatgpt is letting you exercise,” my wife replies. The comment helps me realize that while my robot had not once allocated time to physical exercise. When I ask ChatGPT what I should say to my wife in response, it offers the following response: “Thank you for your understanding. While exercise wasn't explicitly scheduled, I'll make sure to find opportunities for physical activity throughout the day. Your support means a lot to me.”
I’m unable to shake the exercise oversight, which, if this project were to continue on indefinitely, will make me statistically more likely to die at an early age. For some reason, my robot schedules an abundant amount of time for me to breathe deeply and drink a single glass of water for 10 minutes, but ignores the benefits of rigorous physical activity. When I ask why, it blames the “specific context” of my day. I know this to be an excuse, because it is the exact excuse I have used hundreds of times in my life.I perform an interview using the questions that ChatGPT has prepared for me and start wrapping up tasks from the call. I’m not exactly sure what that means, so I turn my recorder off, which is usually all I do at the end of a call. My wife then asks me if I’d like to take the dog on a group walk from 5:30 to 6:30. Unfortunately, I am scheduled to be working on chores and household tasks at that time. ChatGPT asks my wife, through me, if the walk can be rescheduled for a later date. My wife says she will go alone. “That's great to hear!” ChatGPT says, blissfully unaware that it is destroying my marriage.ChatGPT tells me to say, “Thank you for understanding. I appreciate you taking the lead and going for the dog walk alone. It means a lot to me and our furry friend. Enjoy the walk, and we can catch up when you're back!" I do as I am told. My wife stares blankly at me, as if she no longer recognizes her husband, and I can’t blame her. I have become a robot-fueled monster.
I realize it is 3:30 p.m., and ChatGPT has scheduled a 30-minute break. My brief excitement fades after the robot informs me I am to allocate the time as such: 10 minutes of stretching to “increase flexibility,” five minutes of drinking a glass or water and eating fruit or nuts, 10 godforsaken minutes of mindfulness, and five minutes of fresh air. While I perform a prolonged forward fold in the living room, my wife asks me to reiterate the purpose of the assignment, which I can no longer remember. Increasingly concerned about my mental health, she also asks why ChatGPT has not scheduled time for me to exercise. While swallowing my fourth or fifth handful of unsalted peanuts, per ChatGPT’s instruction, I tell my wife I will ask ChatGPT once again about the exercise. Graciously, my benevolent robot says that the lack of exercise was an “oversight” and sends a revised schedule for the remainder of the day, which, inexplicably, includes another immediate one-hour break, followed by 30 minutes of dog walking and 30 minutes of dinner, before I perform chores for two hours, exercise for 30 minutes, and then begin to wind down and relax at 8:30
This revised schedule also does not ask me to work from 4:00 to 5:00 on the feature edits I am now laughably behind on, nor from 7:00 to 9:00. That is a total of three fewer hours of work, and effectively means I am done for the day. The shift seems nonsensical, and it fills me with a joy I haven’t felt since the moment the exercise first started, when ChatGPT allowed me to play with my dog two times in a row. The memory of playing with my dog makes me realize two things: that I once felt joy as a result of this arrangement, and that I have a dog, who, I also suddenly realize, looks as miserable as I am.Any joy I feel escapes my body upon the realization, once again, that a break is no longer a break. This time, ChatGPT schedules a full 15 minutes of meditation and breathing exercises, 25 minutes of self-care, and 20 minutes of hobbying. The prospect of breathing deeply for what feels like the 1,000th time makes my body tense up, so I opt to do that last and choose self-care, which ChatGPT says includes stretching, practicing self-reflection, or journaling. The prospect of participating in any of these three activities again makes my back immediately tense up, so I opt for a fourth option: going for a short walk.I retrieve my dog and go for a short walk, return, and practice 15 minutes of staring into space and 20 minutes of hobbying, my chosen hobby of choice being cleaning up. I then realize that ChatGPT has scheduled me to walk my dog once more for 30 minutes starting at 5 p.m, which is not great planning as my wife is planning to take the dog on the aforementioned 60-minute walk at 5:30 p.m. Worried about my wife losing her mind and increasingly self-conscious about this idiotic exercise, I slink off outside without my wife to complete the task, only to realize 15 minutes in that if I come back at 5:30 p.m., my wife will be late. I run home worried I am failing the two most important presences in my life, my wife and ChatGPT. Back home and my wife none the wiser, she leaves, and I begin 30 minutes of making and eating dinner. While sauteing some kale, I think about all the things I have not had time to do amid the constant back and forth between work and self-care, like see friends or buy my mom a gift for Mother’s Day. Nevertheless, as I begin my two hours of chores, starting with a painting project, I find myself feeling something bordering on relaxation as it dawns on me that I have two uninterrupted hours before a scheduling shift.That brief sense of relaxation, however, is instantly erased upon a call from my wife, who informs me that the group dog walk has gone horribly wrong. At some point on the walk, the group must have stumbled upon a tick infestation, she tells me, as the entire group—and the dogs in particular—have become covered in ticks. By covered, she clarifies, she does not mean one or two each. Rather, she says, she has pulled roughly 50 ticks from our dog, Ziti, and many from herself as well. I suspect this is an exaggeration, albeit a justifiable one, until she returns home and hands me the tick-infested dog while she runs to the shower. Suffice it to say, this is not what ChatGPT had planned, and I spend my entire chore block and then some picking ticks out of my dog, finding some 15 on my own. My wife is beside herself, feeling phantom ticks everywhere. I feel bad for my wife, but I also feel bad that I have disobeyed my robot. “I'm sorry but my wife called me in the middle of painting and informed me that that she and everyone else on her walk got swarmed by ticks. She told me—I shit you not—that she had pulled fifty ticks off the dog and many off herself. I didn't believe her, until she came home and I alone pulled 15 ticks off the dog and a few off her. This has thrown me off schedule. What should I do?” I ask ChatGPT.The robot is at first understanding, telling me to prioritize removing ticks from all relevant parties and adjust my schedule to “accommodate the additional time required for tick removal and potential cleanup.” The robot also emphasizes the importance of practicing self-care during unexpected events. But when I ask ChatGPT for an updated schedule at 7:47, the robot suggests I “take a break, relax, and spend time with your wife” until 8 p.m. before spending a previously unscheduled hour working on a feature, then engaging in self-care and eating a second dinner at 9 p.m., and potentially resuming feature editing from 10 p.m. to 11 p.m. I feel infuriated and ask why the robot has only allocated 13 minutes to spend time with my wife, why the exercise is gone, and why I suddenly have to work for two additional hours. The robot says it’s “understandable that a 13-minute break may not be sufficient for meaningful relaxation and quality time with your wife,” then offers a revised schedule that now requires me to exercise in the dark at 8:30, get back to work at 9, and then have dinner once again at 10. “This is insanity,” I say. “I apologize if the schedule adjustments and suggestions have been overwhelming or unsuitable for your needs,” ChatGPT responds. “It's important to remember that the suggestions provided by an AI model like me are based on general guidelines and may not always align perfectly with individual circumstances or preferences.”
I tell my human boss about my robot boss’ increasingly illogical and borderline abusive demands, and he mercifully gives me permission to end the experiment. Maybe one day the robot boss will be able to understand the emotional minutiae that comes with a dog unexpectedly being covered in dozens of ticks, but today is not that day.I inform ChatGPT that the experiment is over, to which it says:
“If your editor has given you permission to end the experiment, it's important to respect their decision and prioritize your professional responsibilities. You can thank them for their understanding and support, and communicate that you will resume your regular work routine moving forward,” ChatGPT says. “Ending the experiment allows you to regain control over your schedule and work in a way that feels more comfortable and familiar to you. Remember to communicate openly and honestly with your editor, expressing your gratitude for the opportunity and any insights gained from the experience.”I close ChatGPT’s tab on my browser and feel free. For the rest of the night, I engage in little to no self-care. I drink an alcoholic beverage. I stay up to watch the Lakers game. I get out of bed late. I am not a self-caring productivity machine, but I no longer want to be. I just want to control my own life.