AI & Automation
People Use AI Every Day and Don't Trust It. The Brands Pushing It Haven't Caught Up Yet.
79% use AI daily. 76% don't trust it. Brands are misreading the gap.
TL;DR
- 79% of people interact with AI daily. Only 27% think they do. That gap is being read completely wrong.
- 76% of Americans don’t trust AI. They use it anyway. Habit and trust are not the same thing.
- 50% of consumers say they’d prefer brands that don’t use AI in customer-facing content. And brands are doing the exact opposite.
- The moments AI damages brand trust aren’t random. They’re the moments that were supposed to feel personal.
- The feedback loop is slow. By the time it shows in your retention data, it’s already been running for months.
There’s a number worth sitting with before we get into any of this. 79% of people interact with AI-powered tools every single day. Only 27% of them think they do.
The brands seeing that first number are feeling good about it. That’s the wrong read.
The infrastructure problem nobody wants to name
AI stopped announcing itself a while ago. It’s in the spam filter that cleared your inbox, the navigation app that rerouted you around traffic without asking, the bank that flagged a weird charge before you noticed it was there. None of those moments felt like “using AI” to the person experiencing them. They felt like things working, or occasionally not working. That’s what happens when something becomes infrastructure. It disappears into the background and you only notice it when it breaks.
The teams building on this watched it happen and decided it meant people were comfortable. What it actually means is that people stopped noticing it in the places where it stayed invisible. The second it becomes visible, a customer service chat that clearly isn’t a person, a product description that reads like it was generated in four seconds, an email response that refers to their issue in the most generic terms possible, the dynamic shifts. Not dramatically, not with a complaint letter. People just get a little colder. They trust you a little less. They might not even be able to tell you why.
What the trust numbers actually say
76% of Americans don’t trust AI. Not “have some concerns” about it. Don’t trust it. Only 21% say they trust AI most or almost all of the time. Excitement about generative AI specifically dropped 7% since 2024, and 30% fewer people than last year say they expect AI to play a meaningful role in their daily decisions going forward.
These are the same people using it every day, which tells you something important about what daily use actually means. People use the self-checkout because the line is shorter, not because they prefer it. Habit and trust are not the same thing, and the brands that can’t tell the difference are going to keep being surprised by their own data.
56% of users regularly get wrong or useless responses from AI tools and go back anyway. That gets read internally as engagement. It’s not. It’s people not having a better option and making do with what’s there, which is a very fragile position to build a customer relationship on.
The cases where it went visibly wrong
H&R Block and TurboTax’s AI chatbots were providing inaccurate answers more than half the time. That’s not a small error in a low-stakes context. That’s the core product failing at the one thing people came for. The brand trust those two built over decades doesn’t just survive that kind of thing unchanged.
Google ran an Olympics ad showing a father using Gemini to help his daughter write a letter to her athletic hero. The idea was that AI helps you express what you actually feel better. What people saw when they watched it was a father teaching his daughter to outsource the most personal kind of communication to a language model. The backlash wasn’t really about the technology. It was about what the technology implied in that specific moment, that the words didn’t need to actually come from her. Google understood what the tool could do. They didn’t think carefully enough about what using it there would say.
That’s a pattern. The capability gets figured out. The context doesn’t.
What people actually said they want
50% of US consumers say they’d prefer to buy from brands that don’t use generative AI in customer-facing content. Half. And the opposite is exactly what’s happening right now, without much consideration of how that preference is sitting out there.
53% of people say they trust a business less when it relies mostly on AI for customer service. 78% would choose a business where a human answers the phone over one where AI does. 9% would choose the AI. Brands are building toward the thing their customers are actively telling them they don’t want, and the justification is always efficiency, which is a real consideration, just not the only one.
The thing people are actually fine with is AI they can’t see. The fraud detection running quietly in the background, the routing algorithm that keeps their package on schedule. None of that bothers anyone. What bothers people is when AI shows up in a moment that was supposed to feel like someone gave a specific damn about their specific situation. Customer complaints. Renewal conversations. A response to a problem they’ve been patient about. Those are the moments where “we used AI to save time” registers as “you weren’t worth a person’s attention,” and that reading tends to stick.
The social media piece nobody connects to this
Time on social platforms is down roughly 10% since 2022. Algorithm fatigue gets the most airtime as the explanation, and it’s real. What doesn’t get talked about much is that feeds filling up with content that doesn’t feel like it came from an actual person is contributing to that disengagement in a way that’s hard to measure but pretty obvious to observe. People scroll differently when nothing surprises them. They open apps with less intention. They stop expecting to find something worth their time.
The brands pushing the most AI-generated content into those feeds are accelerating the exact thing making their audience less likely to be there.
What the failure pattern looks like internally
Microsoft published research on why AI initiatives fail inside teams and landed on something uncomfortable to read if you’re in a leadership role. The biggest barrier isn’t the technology and it isn’t the employees. It’s the culture. Teams that can’t answer what they’re actually trying to do before they deploy AI, not “increase efficiency” but something more specific than that, tend to produce faster versions of the same unclear outputs they had before.
70% of AI failures trace back to people and process problems, not model problems. The model isn’t the issue. What gets built around it is.
The slow feedback loop
The feedback loop on all of this is slow, which is part of why it keeps happening. People don’t send complaints when they start trusting a brand less. They just change their behavior slightly, engage a little less consistently, take slightly longer to come back. By the time that shows up in retention data or NPS it’s already been running for months, and whatever caused it has probably scaled across the whole team by then.
Questions people keep asking about this
Why don’t people trust AI if they use it every day?
Because using something and trusting it aren’t the same thing. People use AI tools the same way they use a lot of things they’re ambivalent about, because it’s faster, because it’s there, because the friction of the alternative is higher. That’s adaptation, not endorsement.
What do consumers actually want from brands using AI?
To not notice it in the moments that matter. People are fine with AI handling things in the background. They’re not fine with it standing in for a real response when the situation called for one.
What’s the most common mistake brands are making?
Applying AI broadly and calling that a strategy. The brands getting into trouble aren’t always using bad AI. They’re using it in the moments that were holding the customer relationship together, and they’re finding out afterward.
How do you know which moments those are?
Ask which interactions customers actually remember. Not which are most frequent, which ones changed how they felt about the brand. Those are the ones to be careful with.
Sources
- Pew Research Center: Key findings about how Americans view AI
- TechCrunch: As more Americans adopt AI tools, fewer say they can trust the results
- Android Headlines: 32% of consumers use AI every day, but trust remains low
- Klaviyo: Consumer trust in AI, what brands need to know in 2026
- AnswerConnect: AI is losing customer trust and businesses are paying the price
- EIN Presswire: AI trust crisis deepens in 2026
- GeekWire: Microsoft’s AI paradox research
- AIPMM: AI and the rapid rise of brand distrust
- CX Dive: Customers don’t trust AI, and the rift might be hurting business
- Cloud Geometry: What brands will get wrong about AI in 2026
