Exploring the real reasons why some people choose not to use AI

Generative artificial intelligence is everywhere, but not everyone is ready to embrace it—and it’s not just people who fear that AI might replace their jobs or that ChatGPT will become sentient and take over the world.

In a recent study, BYU professors Jacob Steffen and Taylor Wells explored why some people are still reluctant to use generative AI (GenAI) tools. While some people might worry about an AI apocalypse, Steffen and Wells found that most non-users are more concerned with issues like trusting the results, missing the human touch or feeling unsure if GenAI is ethical to use.

“When people actively choose not to use something, there’s often a lot of thought or intention that goes behind it,” said Steffen, a professor of information systems in the Marriott School of Business. “And with the rapid advancement of AI, it’s important to understand why some people are hesitant to adopt such technologies.”

Steffen, Wells and co-authors conducted the study using two different surveys. In the first, participants described situations where they chose not to use GenAI and explained why. The researchers then used those responses to design a second survey, where participants rated how likely they would be to use or avoid GenAI in each scenario. The participants also rated their level of concern over factors such as ethical worries and perceived risk of use.

The surveys revealed common reasons why people chose not to use GenAI. The top four cited concerns were:

1. Output quality: fear that GenAI output is inaccurate or unreliable
2. Ethical implications: fear that GenAI is illegal, dishonest or immoral
3. Risk: fear that GenAI is unsafe and uses private data
4. Human connection: fear that interacting with GenAI is artificial and lacks interactional benefits

These concerns came up across various situations, like using GenAI for schoolwork, writing poems or making art, seeking financial or medical advice, crafting eulogies or heartfelt messages and making important life decisions.

In educational scenarios, Wells said that GenAI can sometimes be a crutch and a replacement for actual learning.

“If you use GenAI for all your assignments, you may get your work done quickly, but you didn’t learn at all,” Wells said. “What’s your value as a graduate if you just off-loaded all your intellectual work to a machine?”

Steffen compared GenAI to a hammer—useful in the right context but unnecessary, or even counterproductive, in others. When used appropriately, Steffen said GenAI can be the most powerful learning tool humanity has ever seen.

Steffen hopes his research will help people make smarter choices about when AI is the right tool to use, as well as foster understanding between avid GenAI users and hesitant non-users.

“It is important to understand your goal behind creating something,” he said. “Is it to learn? Is it to get something done fast? Are you presenting something as your own work, or as a gift to someone else? These questions can help us decide when to use AI.”