I grew up with AI—sort of. As a Millenial/Gen Z cusper, I was almost 13 when Siri was released. She could send a text or call your mom or be tricked into saying naughty words in her funny robot voice, and everyone knew her name. But she wasn’t really a part of our lives in any innate way.
Nowadays, the average teenager lets a cheery AI TikTok voice recount a story so embarrassing you couldn’t financially extort it out of me, then hits post. I had vowed to stay far away from such evils.
But when one of my coworkers posted on Instagram about letting ChatGPT plan her meals for her, I was, admittedly, compelled. As a #busyworkingprofessional, I tend to skip shopping lists and buy, like, canned clam chowder and mixed nuts while hungry at the grocery store.
I imagined that the AI’s capacity to create meal plans based on your desired caloric and macronutrient criteria could help foster a more efficient and healthy me. If AI could plan my meals for me, what else could it do?
I decided to allow AI to plan my life—my meals, workouts, outfits, even my playlists and social life—for five days. Here’s what robots decided would make for the most enlightened Amelia:
The Prep: Setting Up My AI-Generated Week
Food Shopping & Meal Prep
Rather than focusing on calories, I asked ChatGPT for a healthy macro balance for a 25-year-old woman. It suggested that my daily intake be 40-60 percent carbs, 15-25 percent protein, and 25-35 percent fat. Then it repeated stuff like this for 500 words until I clicked “Stop Generating:”
Ignoring the caveats, I asked ChatGTP to create me a five-day vegetarian meal plan with three meals and a snack that met those macro balances. Then I asked for a grocery list based on that plan. My bill at Trader Joe’s was about $73—an extremely normal price for groceries, especially considering that I wouldn’t be spending any money eating out this week.
Exercise Schedule
Next, exercise. I told ChatGPT my age and gender, that I am not trying to lose weight, and that I am “moderately active.” Each day’s workout clocked in at about an hour, and there was plenty of flexibility to choose the direction I’d like to take my exercises.
Making New Friends
Wondering if AI could make me more friends, I decided to set up a Bumble BFF profile. For my photos, I paid $19 to a service called PhotoAI, which promised me the images in its “Tinder package” would be “the best [I’ve] ever looked.”
I uploaded a mix of up-close, three-quarter, and full-body shots with different poses and expressions. I received an email a few hours later with images that ranged in humanness from “me if I used FaceTune” to literal poltergeist. Even the ones that looked, generously, pretty good had too many fingers or slightly disconcerting eyes. I selected six.
For my bio, I fed the AI a few of my interests and specified the app I was using. The prompt took some refining (the first bio it wrote was 160 words long), before offering up a serviceable, if generic, bio:
When it came to profile prompts—quick, personality-driven questions that seem pulled straight from a Tiger Beat interview—ChatGPT appeared determined to portray me as an unassailable dweeb. At first, I was prefacing all my requests with “please,” but as I continued, I started forgoing the formalities and unleashed my inner tyrant.
Finally, I settled on the AI’s suggestion for “Go-to karaoke song,” informing my future BFFs that, “My go-to karaoke song is ‘I Will Survive’ by Gloria Gaynor!/”
Curating a Music Playlist
Spotify recently rolled out an AI feature, a smooth-voiced DJ named Xavier (“My friends call me X!”) who susses out your taste to play old favorites and deliver new music you might like. Every three or four songs, Xavier “switches up the vibe.” Anytime I wanted some tunes in the next five days, I decided, I’d let Xavier call the shots.
Planning My Outfits
I found an app called Style DNA, which uses AI technology to analyze your color, style, and body type in order to suggest the most flattering clothing items and outfits. It provides personalized outfit moodboards and shopping recommendations, and you can upload photos of your own clothes.
After answering questions about my figure and style goals and uploading a selfie, the app determined that I was a “soft autumn” with a “classic gamine” style type. It suggested I look for clothes in shades of rose, olive, avocado, taupe, and teal. My “ideal silhouette,” it said, “is clean with symmetrical, fine lines contrasted with horizontal and geometric cuts,” and I was supposed to wear “textiles of medium density and stiff texture” like wool, cotton, tweed, leather, and denim.
I started individually photographing everything I owned and uploading it to the app for outfit suggestions. The more I uploaded, the better it was able to pair items—while throwing in some shoppable pieces, too. The AI tech wasn’t always great at categorizing the pieces. It decided a wrinkled, khaki linen dress was a “bridal gown,” and assigned all of my hats (and one pair of jeans) as “belts.”
The Assignment
Day One: Sunday
Morning: Breakfast today is both Greek yogurt and oatmeal, which feels like a meal for someone who is teething. I’m in a terrible mood, so I go to put on my trusty “Pop Girlies” playlist before remembering that I’m beholden to Xavier. He treats me to the dulcet tones of Paul Simon’s “The Obvious Child.”
I listen while eating my oatmeal and roasting veggies for the next couple days. In between putting things in the oven, I clean essentially my entire house. “This is making me more productive already!” I think.
Afternoon: “Let’s hear some songs from a playlist you’ve been listening to recently,” Xavier announces. “First up, Taylor Swift.” A sampling of my “Pop Girlies” playlist floats from my phone’s speakers. I am content.
I take a break from my cleaning marathon to make lunch, munching on my side (mixed greens with vinaigrette) while I make a lentil and veggie stir fry. My workout’s focus today is “Cardiovascular Exercise.”
Wearing my girlfriend’s half-marathon finisher t-shirt as a motivational tool, I start jogging around the quiet streets of Kensington. For a few minutes I feel hot and interesting, like a real runner—then my lungs start burning. After five minutes, I settle for a brisk walk while gabbing with a friend on the phone.
Evening: I’m fixing my dinner when I become aware of two critical errors: I did not buy tahini or quinoa. The tahini sauce is easily subbed, but cooking quinoa would take a solid half hour. I add a scoop of brown rice from earlier instead, obliterating the macro balance that ChatGPT so carefully planned.
I launch my Bumble BFF account and spend a few minutes swiping. Loathe to miss any potential matches, I swipe right on pretty much everyone.
Day Two: Monday
Morning: Since I spent all of yesterday alternating between pajamas and workout clothes, I take to StyleDNA to pick out my outfit for the first time. I choose a combo without too many shoppable items, subbing my own crossbody for the app’s suggested $295 bag. I wouldn’t have paired wide-legged pants with such a bulky pullover, but I like this outfit—it makes me feel like I run a farmers market stand.
Breakfast today slaps. It’s tasty, satiating but not heavy, and quick and easy to make (once my roommate tells me how to poach an egg).
My coworkers offer their take on my AI outfit: “Very Silicon Valley.”
I figure if I’m going to hell, I might as well do it properly. I ask ChatGPT to generate a romantic good morning text to send to my girlfriend, Michaela. It feels shady, like I’m Christian de Neuvillette and the AI is a swagless Cyrano de Bergerac.
Afternoon: A text from Michaela comes through while I’m eating my lunch (feta- and-spinach-stuffed sweet potato, roasted brussels sprouts): Who stole Amelia
I decide to meet my two matches on Bumble BFF, so I ask ChatGPT to draw up a short, friendly greeting for them.
Evening: I am ravenous on my drive home from work and inhale my snack (yogurt with nuts and dried apricots) before hitting the gym with Michaela. Today’s focus is strength training. I warm up with a yoga flow while gym bros bench pine trees all around me.
After dinner (tofu and veggie stir-fry; edamame), I change from gym clothes to… more gym clothes, the core of my AI-generated outfit for my friend Kal’s birthday party at Dave & Buster’s.
While the structured jacket helps, I wouldn’t normally pair a tight top with leggings unless I was going to yoga or barre. I swapped the AI’s suggested $151 purple tote for my taupe one and the shoppable sandals (just $26!) for brown heeled ones. The slight heel with workout clothes is objectively silly, but I think the playful contrast is actually the chicest part of this outfit.
Day Three: Tuesday
Morning: I’m working from home today, so I concede to one of StyleDNA’s less-sane outfit ideas. I typically only wear these shorts for Pride, but the app seems to love them—it’s come up with several outfits featuring them. I grew up a stone’s throw from Coachella, so, frankly, boots and cutoffs feel like home to me.
Afternoon: I’m up to seven matches on Bumble BFF. I send more greetings, offer ChatGPT–generated replies to my new friends.
I did not anticipate how embarrassing this whole process would be—I would never, ever interact with humans this way on my own volition, and also my matches seem like cool people I would genuinely like to be friends with, so I regret that I don’t get to talk to them as my actual self.
Evening: I’m on the phone with my brother, complaining about ChatGPT’s determination to make me sound like a loser. “You can ask it to write better answers, you know,” he says, living up, once again, to his reputation as the family tech wizard. “That’s why there’s a chat format. You can say, ‘That’s lame; can you change this part?’”
When a message comes through from one of my Bumble BFF matches, the AI suggests I ask how her day is going—again—I decide to try it out.
Day Four: Wednesday
Morning: I reply to a few more of the friends on my phone. This time, when ChatGPT repeats itself with a ‘How was your day?’ question, it responds to my request for a different query with this, which makes me almost physically recoil with embarrassment:
I copy and paste the message into the chat box, close my eyes, and hit send.
I select my next outfit from the roundup of ideas StyleDNA offers. I love this look. It feels the most “me” of all the AI-generated ensembles so far—in fact, I’ve put together almost the exact same outfit before. I wear it while eating my honestly gorgeous Chat GPT–suggested breakfast.
On my drive to work, Spotify’s AI DJ attempts to expand my music taste. I fall in love with a song called “Nothing Matters” by The Last Dinner Party and am reminded that there is always something new and beautiful to discover. Thanks, AI, for this and my pancakes from earlier; and that’s all.
No reply to my “What makes you smile” message. Probably because ChatGPT’s replies are so (literally) robotic and unhinged, I haven’t gotten responses past the first few messages from anyone besides a woman named Courtney (not her real name), who sweetly tells me about her favorite plants and asks what I like to bake. I don’t bake, but ChatGPT tells her I’m a whiz at chocolate chip cookies and banana bread.
Afternoon: I snack on hummus and carrots. I’m hungry again pretty much instantly, so I heat up my prepped lunch: lentils, brown rice, and sauteed veggies. Then I bravely endure a meeting despite a lentil-induced stomachache.
Evening: I’m starving after a busy evening of errands and chores, and I still haven’t done my workout or made my dinner—a black bean and veggie bowl with brown rice. My stomach is barely recovered from earlier, so the last thing I want to do is eat more legumes and then also exercise.
Faced with this prospect, I break. Fold like a house of cards. I make a grilled cheese sandwich, a sweet and gentle food that has never, ever activated what is potentially undiagnosed IBS, and I go to bed.
Day Five: Thursday
Morning: I haul myself back onto the wagon and fix breakfast according to my meal plan. It’s delicious, but the omelet is massive, so I save the fruit salad as a snack for later. Then I don today’s outfit. My roommate tells me I look like a cartoon character.
Afternoon: ChatGPT calls for a caprese salad and whole grain bread for lunch, but I cut out the middleman and just make a caprese sandwich. While I eat, I check in on my Bumble BFF. I have new matches, so I ask the AI to create a new greeting with a “fun, interesting question.”
Gag, but I’ve pretty much lost my capacity for shame at this point, so I send this message to five people. Then, out of curiosity, I ask ChatGPT to answer its own question. “It’s going to be Harry Potter,” I think to myself.
A message comes through from Courtney: She wants to know my favorite local bakeries. I ask ChatGPT to answer the message, specifically requesting that it cite a bakery from San Diego, and it just makes up one that doesn’t exist. I send it anyway.
Courtney invites me to join her at a bakery. My robot heart melts.
Evening: Dinner is…surprise! More beans. In a bell pepper. Yawn.
ChatGPT would like me to engage in “active rest” or “an outdoor activity … such as hiking, cycling, or playing a sport,” so I set out on a long walk, which is like hiking but the only nature is dogs. At my girlfriend’s house after, I stare deeply into her eyes and say, “I’M FREE.”
The Takeaways
Did letting AI take over make my life easier? No—not for me specifically, anyway.
As evidenced by my chaotic grocery runs, I don’t normally structure my meals and workouts so much, focusing instead on eating intuitively and building movement into my social life with activities like climbing gym hangouts and hot girl walks with my friends. I did come away inspired to eat better breakfasts (ChatGPT is pretty good at breakfast!) and make more grocery lists, but following a rigid meal and exercise schedule was too much, too fast.
But if I were a busy mom trying to streamline meal planning, inputting my family’s dietary restrictions and preferences into ChatGPT and having it spit out a grocery list could be an amazing time saver (hint: it can also organize the list based on aisle). Folks dipping their toe into working out regularly, or aren’t comfortable in a gym, could find the AI’s simple exercise plans super useful.
While StyleDNA isn’t smart enough—yet—to align with my tastes and deliver many outfits I actually like, I appreciated the ease that came with having my ensemble assigned to me each morning. Future improvements to the technology might prompt me to redownload the app. And speaking of intelligent tech—Xavier the AI DJ is awesome. I’ll continue using it to discover new tunes, and I’d love to see Spotify add features like the ability to request a specific, consistent vibe (for occasions when I’m, say, throwing a pool party and don’t feel like curating a playlist).
One thing I’d never recommend using AI for? Human interaction. Just write your own texts, people! And then rest easier at night knowing that, at least for now, not even the smartest chatbot is as good at interpersonal relationships as you are.
Experiment now over, I go to delete my Bumble BFF, but then… I think of Courtney, sweet Courtney who responded to my messages through it all—my soulless photos, my fake bakeries, my exclamation point overuse. I can’t just disappear on her.
I write her a message with my own human brain, explaining that I’ve been a robot this whole time. I tell her that I understand if she never wants to speak to me again, but I would really, really like to be friends, because she seems like a good person who didn’t deserve to be on the wrong end of my technological experiment. And then I wait.
I can only hope she offers me something distinctly human: grace.
