I didn't choose to start using Meta AI. It arrived — in Instagram's search bar, pinned at the top of a WhatsApp chat list, sat inside Facebook's own search results — and I spent about six weeks tolerating it before I sat down and switched it off, app by app, on purpose. This is the honest version of why, not the polite one.
None of this is a grudge against the model itself. I use ChatGPT and Claude most days for actual work, and I don't think Meta's assistant is meaningfully worse at answering questions than either of them. The problem was never the quality of the answers. It was that I never asked for it to be there in the first place, and once I started paying attention, I noticed how much of my daily app use had quietly gained a third participant.
The moment it stopped feeling optional
The tipping point was small and specific: I opened WhatsApp to reply to my sister and found a Meta AI thread pinned above every real conversation I had, with a little "Ask Meta AI" prompt sitting where the most recent message preview used to be. Nobody added it. It was just there, in the same visual slot as the people I actually talk to, competing for the same half-second of attention.
Instagram did the same thing from a different angle — the search bar stopped being a place to look someone up and became a place where Meta AI offered to answer a question I hadn't asked, every single time I tapped it. Facebook's version was quieter, tucked into the search results page, but it followed the same pattern: an assistant inserted into a spot I already used for something else, betting that proximity would be enough to make me try it.
What I actually used it for, honestly
I gave it a fair go before deciding any of this. I asked it to summarise a long group chat I'd fallen behind on, to suggest a caption for a photo, to answer a quick factual question while I was scrolling. It did all of that competently. None of it was bad. But every single one of those tasks is something I already do better in a dedicated AI app, on purpose, when I've actually opened it to think something through — the same discipline I've written about before when it comes to choosing which AI tool earns a place in a workflow instead of just taking whichever one showed up first.
That's really the whole case against it in one sentence: Meta AI never had to win me over, because it never had to be chosen. It just had to be there often enough that using it eventually felt like the path of least resistance.
"It wasn't a worse assistant. It was an uninvited one, sitting exactly where I go to talk to actual people."
Where it changed how the apps themselves behaved
The part that actually pushed me to switch it off wasn't Meta AI answering questions — it was what its presence did to the apps around it. WhatsApp's chat list is a list of people. Adding a pinned AI thread at the top of it, in the same row style as a real conversation, blurs a line that used to be simple: this app is for messaging humans. Instagram's search bar used to be a single, boring, reliable tool for finding an account. Now it's also a chat window, and it's not always obvious which mode you're in until you've already started typing.
None of that is dramatic. It's the kind of change that's easy to shrug off individually and only adds up when you notice it happening in three apps at once. It reminded me of the same pattern I went through with every notification I'd let apps default to without ever deciding I wanted them — small, unrequested defaults that quietly become normal simply because nobody made you choose them.
What I actually turned off, app by app
There's no single switch, which is its own small frustration. I muted and archived the pinned Meta AI thread in WhatsApp so it stopped sitting above real conversations. In Instagram, I turned off the AI search suggestions in the app's settings so the search bar went back to being a search bar. In Facebook, I hid the Meta AI shortcut from search results the same way. Each one took under a minute once I found the right settings screen, and each one had to be done separately — there's no account-wide toggle that clears it from everywhere at once.
The honest verdict: Meta AI isn't a bad assistant, and I'd say the same about most default-on AI features arriving in apps you already use. The actual issue is consent — being opted into an assistant living inside your messaging and search tools rather than being asked. Once I turned it off in each app, nothing I actually wanted to do got harder. That's usually the tell that it wasn't earning its place.
What I use instead, and why the separation matters
I still use AI constantly — more than most people I know. It's just confined to apps I open specifically to use it, where I'm choosing the tool for the task rather than accepting whichever one Meta decided to surface that week. If you want to see what a topic hub of that kind of deliberate tool-picking looks like, the AI Tools section is where I keep the honest comparisons and workflows I've actually kept using.
It's exactly why Samsung launching its own Gemini-powered glasses to rival Meta's didn't shift anything for me — a smarter assistant under the hood doesn't fix an assistant that keeps volunteering itself, which was always the actual complaint.
The line I've landed on is simple: an AI assistant is welcome in my life when I open it on purpose. It's not welcome pinned above a message from my sister. If you want the next honest verdict on something like this before it goes live, join the Brite newsletter.
Questions people actually ask
Is it safe to leave Meta AI switched on in WhatsApp?
Nothing about it struck me as unsafe, exactly — it's more that it changes what a chat thread is for. Meta AI sits in your inbox like a contact you never added, ready to answer questions inside a conversation you thought was just you and the person you were messaging. That's a reasonable feature for some people. For me it made WhatsApp feel less like a private line and more like an app with a third party quietly present, which is the thing I actually turned off.
Can you fully delete Meta AI, or only hide it?
In every app I checked, it's hide rather than delete. You can remove the shortcut, mute the chat and stop it responding to your messages, but the underlying assistant is still part of the app — it just stops surfacing itself. That distinction mattered to me more than I expected: turning it off felt like closing a door, not like it had left the building.
Does turning off Meta AI change your Instagram or Facebook recommendations?
Not that I noticed. My Explore page and main feed kept behaving the way they always have — same mix of accounts I follow and the algorithm's usual guesses. Meta AI's search assistant and the underlying recommendation engine appear to be separate systems, so switching off the chat assistant didn't visibly touch what showed up in my feed.
Is Meta AI the same tool across Instagram, WhatsApp and Facebook?
Same underlying assistant, but it shows up differently in each app and has to be switched off separately in each one — there's no single account-level toggle that clears it everywhere at once. I turned it off in Instagram's search bar, muted the pinned chat in WhatsApp, and hid the shortcut in Facebook, and each of those was its own small settings hunt.
What do you actually use instead of Meta AI?
ChatGPT and Claude, in a separate app, opened on purpose when I actually want something answered. That's the real difference for me — it's not that Meta AI is a worse model, it's that I don't want an AI assistant living inside the apps I use to talk to people. Keeping the tool and the conversation in separate places turned out to matter more than which assistant I picked.