Google's latest Pixel update was supposed to make your phone smarter.
Instead, a growing number of users say it's confidently inventing details about their jobs, their families, and their health—then presenting the fiction as fact.
Welcome to the newest chapter in the uneasy marriage between artificial intelligence and the device in your pocket.
The complaints center on Gemini's habit of summarizing personal context it doesn't actually have.
Pixel users report the assistant drafting emails that reference meetings that never happened, suggesting replies to messages from contacts who don't exist, and generating "recaps" of conversations the user never had.
One user described a summary of a family group chat that invented an argument, complete with fake apologies.
What makes this different from the usual chatbot hiccup is where it happens.
It's woven into the keyboard, the messaging app, the photo editor, and the notification shade.
When Google positioned the Pixel as "the AI phone," it bet that convenience would outweigh discomfort.
That bet looks shakier every time the assistant fabricates a detail inside a text you were about to send to your boss.
You trust it with your calendar, your passwords, your kid's school emails.
Once you catch it inventing something, you start double-checking everything—which is exactly the mental overhead a "smart" assistant was supposed to remove.
Users on forums describe disabling features not because they're bad, but because they can't predict when they'll be wrong.
Google's response has been the industry-standard shrug: these are known limitations of large language models, accuracy is improving, and users should verify important information.
That's a reasonable engineering answer and a terrible consumer one.
Nobody buys a $799 phone expecting to fact-check their own text messages.
The deeper issue is philosophical, and it's landing in American kitchens and commutes right now.
We've spent a decade teaching people to outsource memory to their devices—phone numbers, directions, birthdays.
Now the device is outsourcing its answers to a system that guesses.
The result is a population that trusts technology more than it trusts itself, and a technology that hasn't earned it.
The hardware is genuinely good, the camera is excellent, and some AI features work quietly and well.
But the gap between Google's marketing and Google's reality is widening, and customers are the ones standing in it.
A phone that lies to you isn't a smarter phone.
The uncomfortable truth is that we asked for this.
We wanted assistants that anticipate, summarize, and speak in full sentences.
We got them, and they occasionally make things up.
Final Thoughts
The question isn't whether Google can fix it—it's whether we'll keep rewarding companies that ship the illusion of intelligence before the substance.