Mental health chatbots face a cultural divide over emoji use and conversation depth

Artificial intelligence chatbots are changing how people seek help for anxiety, depression, and other emotional struggles. A recent study found that while Americans prefer ongoing conversations and dislike when automated systems use emojis, Chinese users easily accept the technology regardless of its communication style. The findings were published in Computers in Human Behavior.

People often feel comfortable discussing sensitive information with digital tools because the interaction offers anonymity and immediate responses. Even general platforms that are not meant for medical use are becoming everyday outlets for people to share their daily stress or deeper psychological concerns.

Because so many individuals are turning to these digital tools, researchers are trying to understand what makes a non-human conversation feel helpful. Text-based exchanges lack the physical cues of normal speech, such as a sympathetic tone of voice or a warm facial expression.

Emojis are digital images that try to bridge this gap by adding emotional tone to text. While a smiley face might make a casual text seem friendly, it can also make professional advice seem less credible. In high-stakes situations involving personal well-being, users might demand a more serious tone.

Another factor is the way a person asks for help, which psychologists link to different ways of coping with stress. Emotion-focused coping happens when a person vents or shares deep personal struggles to regulate their negative feelings. Problem-focused coping happens when a person asks concrete questions to gather facts and find a direct solution.

Free daily newsletter

Finally, an automated system can maintain different levels of true conversation. Some interactions are just a single turn, where the user asks a question and the system gives one answer. Others are multi-turn dialogues, where the system remembers context and responds to follow-up questions over a longer time horizon.

Jihye Lee, a communication researcher at the University of Texas at Austin, led a study to see how these three factors alter user trust. The research team included Zinan Darren Yang and Weijia Shi of the University of Texas at Austin, along with Yan Liu of Shanghai University.

They set up an online experiment designed to compare reactions across two distinct cultural settings. The team recruited 394 participants from the United States and 401 participants from China.

The researchers wrote simulated transcripts featuring a human talking to ChatGPT about mental health concerns. To make the conversations realistic, the team sourced the initial human messages from actual posts on Reddit and a popular Chinese support forum.

The team tweaked the transcripts in three ways before showing them to participants. First, they instructed the system to either include emojis in its replies or to use plain text.

Second, they changed the type of user prompt to reflect the two different coping styles. Half of the transcripts showed a human sharing vulnerable emotions. The other half showed a human asking direct questions about dealing with anxiety and depression.

Third, the researchers varied the depth of the conversation. Some participants read a single exchange between the human and the machine. Others read an extended dialogue where the human asked follow-up questions and the machine offered step-by-step guidance based on previous answers.

After reading the assigned transcript, each participant answered a survey. They rated the quality of the information, the emotional support they felt the machine provided, and their likelihood of using the system themselves.

The results revealed major differences in how users from the two countries evaluated the interactions. For the American participants, the presence of emojis had a noticeably negative impact.

When the machine used emojis, American users gave the responses lower ratings for information quality. Emojis also harmed the machine’s reputation when the human in the transcript shared deep emotional struggles. In those highly emotional scenarios, the use of emojis made Americans less likely to say they would use the service.

However, interactivity proved to be a highly positive feature for the American group. Transcripts that showed a sustained, back-and-forth conversation boosted user ratings across all categories.

American participants felt the extended dialogues contained higher quality information and offered better psychological comfort. The prolonged conversational style also increased their willingness to recommend the tool to a friend in need.

Reactions from the Chinese participants told a totally different story. Their survey scores remained completely stable no matter how the transcripts were altered.

The Chinese users did not care if the machine used emojis, what kind of prompt initiated the conversation, or how many turns the dialogue took. Out of all the groups tested, the Chinese participants consistently reported the most favorable views of the technology.

They also showed the strongest overall intention to use the digital tools for their own health needs. The researchers proposed that Chinese users might interpret digital interactions based on the entire context rather than small stylistic details like a digital frowny face.

China also heavily integrates automated services into its national infrastructure and daily life. This high level of societal integration might give Chinese users a stronger baseline of trust in digital tools compared to American users.

The authors noted a few limitations to their research. They chose to use a general conversational model rather than a clinical tool that has passed rigorous medical trials.

The experiment only measured subjective opinions about the conversation. It did not test whether talking to the machine actually reduced symptoms of depression or anxiety over a long period.

The team also encountered a translation hurdle when preparing the study materials. The artificial intelligence naturally provided longer, more detailed responses in Chinese than it did in English.

To keep the transcripts comparable in length and meaning, the researchers translated the English responses into Chinese using ChatGPT and then had three bilingual native Mandarin speakers review the translations to verify semantic accuracy and contextual relevance. Future studies will need to find new ways to ensure exact linguistic matching without losing cultural nuances.

Because the Chinese participants already held such a high baseline level of trust, their survey answers often hit the very top of the measurement scales. This high starting point may have made it difficult for the researchers to detect minor shifts in opinion.

In future projects, the research team hopes to test if the sheer number of emojis makes a difference. Using a single digital smile might be acceptable, but using a dozen might aggressively damage the system’s credibility.

The current findings suggest that technology companies designing health applications should avoid a universal approach. The communication styles programmed into these services may need to adapt specifically to the cultural expectations of the user.

The study, “AI chatbots in mental Health: How emojis, prompt type, and interactivity shape user perceptions in the United States and China,” was authored by Jihye Lee, Zinan Darren Yang, Weijia Shi, and Yan Liu.