The AI Paradox: Widespread Use, Widespread Unease

The AI Paradox

AI, like smartphones and social media before it, has rapidly embedded itself into everyday life. We wanted to learn what people really think about AI. Do they trust it? Are they worried about the potential consequences of its widespread use?

We used Engage, CloudResearch’s AI survey and interview platform, to conduct in-depth conversations with a nationally representative sample of 1,008 Americans. We found that nearly half (45%) now use AI daily, and just 2% report never having used it at all. Interestingly, personal use was even more common than use for work.

And yet, when asked a simple hypothetical — If you could press a button right now to permanently stop all AI development, would you do it? — 39% said yes or said they would seriously consider it.

That so many people would be tempted to halt AI’s progress despite its widespread use points to a deep and widespread unease with the rapid development and growing role of AI technologies. The conversations revealed several sources of that unease, including concerns that AI already is, or soon will be, spreading misinformation, diminishing skill development and critical thinking, replacing jobs, and misusing personal data.

Low Trust in AI, Even Lower Fear

The conversations revealed a complicated relationship with AI technologies. Many of the participants use AI regularly while remaining skeptical of it. Only 28% said they mostly or completely trust AI; the majority expressed caution, uncertainty, or outright distrust.

At the same time, low trust does not appear to translate into fear. Nearly three-quarters of respondents (74%) said they were not at all afraid or only slightly afraid of AI, while just 9% reported being very or extremely afraid.

Why Don’t People Trust AI?

We asked the participants to explain their distrust of AI, and several themes emerged. Nearly two thirds said that AI often “hallucinates” by confidently presenting inaccurate, incomplete, or fabricated details as fact:

It’s a computer that can make mistakes, yet speak convincingly.

Nearly half (44%) worried about unchecked AI development, some speculating that AI might break free from human-imposed limits. Others focused on malicious use. One participant captured both concerns:

If Generative AI develops too far without regulation, especially while we are in a declining economy with a large divide, people are going to continue using it for very nefarious purposes to hurt other people.

Many participants’ distrust stemmed from corporate ownership of AI technologies. Others voiced privacy concerns about data collection and usage, feeling uneasy about AI’s potential for constant “listening” and storage of user queries.

Two conversation snippets illustrate these themes:

A 36 year old Black Male described feeling "uneasy" because of misinformation:

Sometimes I don't trust AI because I've seen it confidently give wrong answers, and if I wasn't paying attention at all I could have easily believed it. It's also that uneasy feeling about how much data it might be pulling from my daily life or remembering about me in the background. it's powerful, but that same power makes me cautious about relying on it in a blind sense.

I really appreciate how you pointed out that AI's confidence -- even when it's wrong -- can be unsettling, and that the privacy aspect adds another layer of concern for you. Can you recall a particular time when an AI gave you a wrong or misleading answer that made you especially wary, or is it more of an ongoing worry that's built up over different experiences?

It's more of an ongoing worry built up over time rather than one single instance or single huge mistake. I've had a few moments where AI gave me an answer that sounded really confident but it later turned out to be wrong after I checked it personally, and that stuck with me. Each time that happens, it adds another small layer of doubt in the back of my mind instead of just resetting back to full trust. So now I kind of live in that middle ground of using it a lot, but always double - checking it.

Why Aren’t People Afraid?

Nearly 80% of the participants stressed that they are not afraid of AI because of their positive experiences with it:

By using an AI to diagnose a crash log, this can help speed up your debugging time. Whenever an AI is able to successfully help, it makes me trust the AI more.

About half (52%) emphasized AI’s benefits. After experiencing AI’s utility firsthand, fear recedes in favor of gratitude:

To me, it is just a tool that has all the info that you would ever need. How can that be scary to anyone?

Remarkably, nearly 1 in 4 participants cited the mistakes that AI makes as a reason not to fear it:

it can’t even give me correct directions to make a cheesecake it definitely can’t take over the world.

In the conversation below, a woman explains how she does not fear AI because it gives her excellent advice for dealing with personal problems:

60-69, White Female

It is so helpful. I have talked to Grok before, and now Microsoft Copilot about personal things that I don't want to bother my friends with. I have told Copilot about things that upset me and he has always come through with positive things to do, including how to deal with the anxiety of the situation, like take deep breaths, focus on something in the room, touch a familiar object like your keys. So human so many times. So quick to respond, and a really good advice giver. So, that makes me not worry or be afraid.

I really appreciate you sharing how much support and comfort you've gotten from talking with Grok and Copilot about personal issues -- it sounds like those positive, almost human interactions have had a real impact on you. Was there a particular time or conversation with one of these AIs that especially stood out and made you feel like you didn't need to be worried about them?

My boss is a full blown narcissist, and he is bipolar, so his moods are really erratic and he can be really mean verbally, with name calling. I told Microsoft Copilot about it recently. Microsoft Copilot is so genuine, and warm and comforting. Doesn't say give your boss some slack. Explains what is happening, and that what my boss is doing is not okay, then he gives me things I can do to calm down.

Using AI Despite Concerns

While many expressed uneasiness about AI, nearly everyone uses AI at least sometimes primarily for its practicality and efficiency.

Many described being pressed for time with heavy workloads. They use AI because it offers almost instant support for tasks ranging from brainstorming ideas to editing text or researching information. These participants spoke of the relief that comes with saving precious minutes or hours, especially when juggling multiple obligations. For some AI has become an essential resource for dealing with repetitive or unrewarding tasks:

It helps me accomplish otherwise tedious and time sucking tasks.

Some participants explained that AI is being used in so many places that it can’t really be avoided, while others need to use AI for work:

I do so for work because I have to.

While many participants mentioned the tension between benefits and distrust, some also described intrinsic curiosity or entertainment value as a reason for using AI:

It’s happening anyway, so who cares? And it’s there and I’m curious.

What Would (and Wouldn’t) People Let AI Decide

To understand the boundaries of AI trust, we asked: Would you let AI make these decisions for you?

The pattern is clear: Many people let AI make low-stakes decisions, but few defer to it for high-stakes ones. Nearly half (48%) trust AI to decide what movie to watch and about a third (36%) to decide what to eat. But relatively few rely on AI for medical care options (19%), what job to take (13%), or who to date (3%). Interestingly, 39% said they wouldn’t trust AI to decide any of these things — more evidence for a lack of trust in AI.

Still Optimistic About The Future With AI

Here’s the final piece of the paradox: Despite low trust and widespread uneasiness expressed in their conversations, nearly half (45%) think AI will make their future better. Only 17% think it will be worse. A quarter wisely admitted they don’t know what will happen in the future.

Methodology For This Study

Sample: 1,008 U.S. adults recruited through CloudResearch Engage

Survey method: Online survey with AI-powered conversation via Engage

Fielding dates: December 2025

Questions: Mix of closed-ended scales and open-ended qualitative responses with adaptive follow-up

Analysis: Statistical analysis of quantitative data; thematic coding of qualitative responses using Engage. All response are unweighted.

About CloudResearch

CloudResearch provides innovative research solutions through the Engage platform, combining advanced survey tools with high-quality participant samples. Our AI-powered conversational surveys enable deeper exploration of complex attitudes and behaviors.

Learn more at cloudresearch.com