Now, this is actually creepy. Last month, I was lying in bed at midnight, thinking about maybe, possibly, buying new running shoes. I did not say this out loud. I did not type it anywhere. I just thought it. The next morning, I open my shopping app and there they are—three different brands of running shoes, in my size, in colours I like, with a notification saying “Ready to start your fitness journey?” I nearly dropped my phone in my congee. Are they reading my mind now? Is that where we are in 2026?
This is AI-powered personalization in Shanghai today. It is not just about remembering what you bought before. That is old news. Now these systems are like that one friend who finishes your sentences, except they finish your thoughts before you even have them. And honestly? Sometimes it is amazing. Sometimes it is terrifying. Sometimes it is both at 7 AM when you are too sleepy to know the difference.
Let me tell you what is really happening with these prediction machines, because I have spent way too much time trying to understand why my phone knows me better than my own mother.
When Your Apps Know Your Life Better Than You Do
Here is a true story. Two weeks ago, I got a notification from my music app. It said “Feeling stressed about your presentation tomorrow? Try this playlist.” I did have a presentation. I was stressed. But I never told the app about it. It saw that I had been searching for “how to not panic during public speaking” at 2 AM. It noticed I set five alarms for the morning. It probably heard me practicing my slides out loud because I forgot to close the other apps. Then it made a playlist of calming piano music and rain sounds. I listened to it. I cried a little. The presentation went fine.
This is what we are dealing with now. The AI does not just look at one thing you did. It connects everything. Your calendar, your search history, your location, how fast you are walking, even how you are holding your phone. My health app and my music app talk to each other now. If I walk more than 10,000 steps, I get upbeat songs. If I am lying in bed all day, it sends lo-fi beats to study to, even though I am not studying, just avoiding responsibilities.
The food delivery apps are the worst best example. Meituan knows my life schedule better than I do. On Monday mornings, when I am sad about the weekend ending, it suggests comfort food—hot and sour soup, fried dumplings, extra spicy noodles. On Friday nights, it knows I want to celebrate with bubble tea and cake. On Sunday evenings, it offers me meal prep options because it knows I will feel guilty about eating badly all weekend. The app is basically my nutritionist, therapist, and enabler all in one.
And it is not just guessing. Last month, I tried to trick it. I searched for salad recipes for three days straight. I clicked on healthy food blogs. I even ordered one sad salad on Tuesday. The app adjusted. It started showing me “guilt-free” versions of my favourite foods. Skinny hot pot. Low-calorie dumplings. It knew I wanted to be healthy but also knew I would not actually be healthy. So it found a middle ground. I felt seen. I felt attacked. I ordered the skinny hot pot.
Shopping That Creates a Version of You That Does Not Exist
Taobao and Xiaohongshu, they are not selling products anymore. They are selling identities. And they use AI to figure out which identity you want this week.
Here is how it works. You buy one thing—just one thing—and the AI builds a whole story about who you are. I bought a yoga mat because my back hurt from sitting too much. Suddenly, my feed was full of athletic wear, meditation apps, organic protein powder, and those expensive water bottles that keep drinks cold for 24 hours. The AI decided I was a wellness girl now. A green juice person. Someone who wakes up at 6 AM for sunrise yoga.
I am not that person. I used the yoga mat twice, then used it as a place to nap while watching TV. But the AI did not give up. It adjusted. It started showing me “cosy wellness”—weighted blankets, aromatherapy candles, comfortable clothes for “restorative rest.” It pivoted from active wellness to passive wellness. It found a way to sell me things that matched who I actually am, not who I pretended to be.
The scariest part? It was right. I bought the blanket. I bought the candles. I am typing this while wrapped in that blanket, smelling like lavender, feeling very seen and very broke.
My friend had the opposite problem. She bought one leather jacket for a costume party. One jacket. Her apps decided she was a “edgy motorcycle girl” now. For months, she got ads for biker boots, skull jewellery, and tattoo parlours. She is a kindergarten teacher. She drives a small electric car. She has never touched a motorcycle. But the AI was committed to this version of her. It took six months of searching for floral dresses and soft cardigans to convince the algorithms that the leather jacket was a mistake.
Entertainment That Writes Itself Around Your Mood
Video apps like iQiyi and Bilibili, they do not just recommend shows anymore. They edit the recommendations based on your current emotional state. I do not know how they know my emotional state. Maybe it is how fast I scroll. Maybe it is the time of day. Maybe my phone camera is watching my face while I browse. I try not to think about it too hard.
Last Wednesday, I had a bad day at work. I opened my video app to distract myself. Instead of my usual crime dramas, it showed me a list of “comfort shows”—old sitcoms, cooking competitions, nature documentaries with soft narration. At the top was a note: “No intense dramas today. You need something gentle.” I watched a show about people restoring old furniture for three hours. It was exactly what I needed. I did not know I needed it. But the AI knew.
Music apps do the same thing but faster. NetEase Cloud Music has this feature where it creates daily playlists based on “your current life phase.” Not just your listening history, but your phase. When I was going through a breakup last year, it gave me sad Chinese ballads, angry rock music, and then slowly, week by week, it introduced more hopeful songs. It was like a friend who knows when to let you cry and when to tell you to move on. Except it is not a friend. It is code. Very empathetic code.
Even the news apps are personalized now. Toutiao does not show everyone the same stories. It learns what kind of news upsets you, what kind bores you, what kind makes you click. I get a lot of stories about local food festivals, tech innovations, and cats doing funny things. My colleague gets politics, finance, and serious international news. We both use the same app. We live in the same city. But we see completely different versions of the world. Sometimes I wonder what I am missing. Sometimes I am grateful to miss it.
The Convenience Trap and the Price We Pay
Let us be real about why this works. It is so easy. The AI removes the thinking part of life. What should I eat? The app knows. What should I watch? The app knows. What should I buy, wear, listen to, read? The app knows. I do not have to make decisions anymore. I just have to say yes to the suggestions.
Last weekend, I tried an experiment. I turned off all personalization features. I cleared my history. I told the apps to stop learning about me. It was chaos. The food app suggested restaurants across the city that would take two hours to reach. The shopping app showed me things I would never wear—business suits, fishing equipment, baby clothes. I have no baby. The music app played random genres that gave me a headache. I lasted six hours before turning everything back on. Six hours of freedom, then I ran back to my AI comfort zone.
But here is what worries me. If the AI always gives me what I want, how do I find new things? Real new things, not the “new” things that fit my old patterns. Last year, I discovered I love jazz music. Not because an app suggested it, but because I walked past a live bar in Jing’an and heard it through the window. The AI would never have suggested jazz. I only listened to pop and indie music. There was no data point connecting me to jazz. I had to find it by accident, by being bored and walking somewhere random.
Are we losing those accidents? Those moments of finding something completely wrong that turns out to be right? I do not know. I hope not.
Living With the Prediction Machines
So here is my life with AI personalization in 2026. My morning starts with my alarm app choosing the perfect wake-up sound based on my sleep quality. My weather app suggests what to wear, considering my calendar and walking distance. My food app orders my breakfast before I am fully awake because it knows I am too sleepy to choose. My commute playlist matches my energy level for the day. My work apps prioritize my emails based on what I actually care about. My evening entertainment is ready before I finish dinner.
Everything is smooth. Everything fits. Nothing surprises me, except when the AI gets it wrong, and then it is annoying because I am used to it being right.
Is this good? I think so, mostly. I save time. I have less decision fatigue. I discover things I like that I might have missed. But sometimes, late at night, I wonder who I would be without all these suggestions. What would I choose if I had to choose alone? Would I be different? Would I be more me, or less me?
The AI predicts what I want before I want it. But maybe wanting is the important part. Maybe the wanting, the searching, the not knowing—that is where we find out who we are.
I will keep using the apps, of course. The blanket is too comfortable to give up. But sometimes, on weekends, I leave my phone at home and walk without direction. Just to see what I find when nobody is predicting anything.
Do you feel like your apps know you too well? Or do you love the convenience? Tell me your weirdest AI prediction story—mine involves being suggested wedding dresses when I was just looking for a birthday gift for my sister.