How We Use AI Chatbots: Habits, Preferences, and Concerns

Every year, AI chatbots attract more and more users, becoming an essential part of our daily lives. At the same time, user habits and preferences continue to evolve. Some people interact with AI daily, delegating the bulk of their tasks, while others remain skeptical and avoid using these tools. A recent survey of Americans conducted by ChatOn explored how the use of AI chatbots has changed in 2025 and what new trends have emerged.

How Skilled Are AI Users?

The first aspect studied in the research was users’ level of expertise with AI chatbots. Respondents were asked to indicate their level of knowledge, from beginner, described as “I’m new to AI chatbots and still figuring out how to use them,” to expert, described as “I have a deep understanding of how AI chatbots work and often create workflows, automations, or integrations involving AI.”

Here’s how respondents rated themselves. Nearly half of them (49%) consider their level of AI knowledge to be intermediate (“I can use AI chatbots confidently for several kinds of tasks”). Only 24% of respondents say they are advanced (“I understand how to get high-quality results and often customize prompts or settings”), and 3% identify as experts. However, about 20% describe themselves as casual users (“I use AI chatbots occasionally for simple or guided tasks”), while just 4% classify themselves as beginners.

How Often Do People Use AI Chatbots? 

The second aspect of the study focused on how frequently Americans use AI chatbots. About 22% use them multiple times per day, 14% once a day, 36% a few times per week, 16% a few times per month, and the rest interact with chatbots even less often. 

With so many AI chatbots available in the App Store and Play Store, ChatOn also looked at how many AI chatbots the average user interacts with. The results showed that 60% of Americans use only one AI chatbot, while 40% use multiple chatbots.

What People Use AI Chatbots For 

One of the topics ChatOn researches is which daily tasks people most commonly delegate to AI chatbots. The most popular task is getting answers or searching for information (74% of respondents). Writing or editing emails, messages, and texts comes in second (65%), followed by brainstorming ideas (54%). 

More than half of respondents also use AI chatbots for tasks such as image generation or editing for fun or non-work purposes (52%), learning new skills or concepts (51%), and summarizing articles or documents (51%). Among less popular tasks, respondents mentioned data analysis (28%), homework or study help (26%), and talking to AI as a conversation partner or for emotional support (21%).

How Users Improve Their AI Skills

The ChatOn survey explored how users improve their skills in interacting with AI chatbots. Nearly half of respondents (46%) experiment on their own with different prompts, 27% ask friends, colleagues, or classmates for advice or examples, 21% watch online tutorials, including videos on YouTube, about prompt engineering or AI assistants, and 16% follow tips or discussions on social media (X, Reddit, LinkedIn, Facebook, etc.). At the same time, 29% don’t actively look for new ways to improve—they just use AI chatbots as best they can.

How Users Prefer to Interact with AI 

The survey also shed light on how users prefer to interact with AI chatbots. 38% of Americans like to type short, direct questions or prompts, 20% experiment with phrasing, formatting, or longer prompts to improve results, and 18% prefer speaking to AI chatbots using voice commands or conversation. Only 12% give context and instructions to get better responses, and 5% copy example prompts they’ve seen online.

5 Common Habits When Using AI Chatbots

The research highlights the most common habits people have when using AI chatbots. Here are the top five: 

1. Asking follow-up or clarifying questions to double-check answers (48%) 
2. Rephrasing the same question in different ways to get better answers (42%) 
3. Verifying AI-generated information using Google or other sources (39%) 
4. Clicking on sources or links that the AI provides (37%) 
5. Using follow-up questions suggested by the AI (36%) 

An interesting additional insight: 28% of respondents compare answers from different AI tools or models.

Privacy, Safety, and Trust Issues 

Nowadays, privacy and safety concerns around AI are among the most debated issues, and research shows how they impact users’ behavior. About 54% of respondents avoid sharing sensitive personal information (e.g., name, financial details, health data), 42% avoid uploading documents or images containing private or confidential data, 36% avoid sharing confidential work-related information, 29%  avoid connecting their AI accounts with other services (e.g., Google Drive, email), and 28% review and adjust privacy settings in their accounts. In addition, 15% use temporary chat or incognito mode when it’s available.

AI Hallucinations: How Often Do They Happen? 

AI hallucinations were another hot topic to discuss. On social media, users frequently share stories of AI chatbots providing incorrect information, sometimes leading to negative consequences, such as a ruined holiday or lost money. Survey respondents, in turn, were asked to rate how often they encounter “hallucinations” when using AI, from “Never” to “Very often,” and the most common answer was “Sometimes”. 

Let’s take a closer look at the details. For each type of issue, users most often reported experiencing it “Sometimes.” About 39% of respondents sometimes received responses that didn’t match their questions, 36% sometimes got outdated information, 33% sometimes noticed contradictory answers to similar questions, and 19% sometimes encountered made-up sources or references. 

Overall, only 5-9% of respondents had never experienced any of the first three types of AI hallucinations. Almost one in four (24%) hadn’t noticed “made-up sources or references.”

What Does It All Mean? 

The survey results indicate that AI chatbots have become an important part of our daily lives, helping users accomplish a lot of tasks—from searching for information to study or work assistance. Most respondents use chatbots weekly and actively look for ways to improve their AI skills. 

What’s important is that most respondents are aware of the privacy and safety concerns associated with AI usage. As a result, they adopt strategies to feel safer— users avoid sharing sensitive personal information or uploading documents and images containing private or confidential data. In pursuit of accuracy, many also tend to ask follow-up or clarifying questions to double-check answers and verify AI-generated content using Google or other sources.

About the Study 

The survey was conducted by ChatOn, one of the most popular AI chatbots with over 90 million downloads and powered by multiple large language models (LLMs). A total of 300 Americans participated in the study conducted in November 2025.