Artificial Intelligence Research: About 30% of American teenagers use AI chatbots daily, with safety concerns increasing.

2025-12-14 06:40:26

Pew Research Center recently released its latest research report analyzing the current status of American teenagers using social media and AI chatbots. The survey covers various indicators such as age, ethnicity, usage frequency, and internet time. Against the backdrop of increasing global concern over online safety for teenagers, this report has once again sparked discussions across different sectors about how technology influences the mental health, dependency, and risk of addiction among young generations.

As countries begin to address related issues, the Australian government has taken the lead by announcing a ban on social media for minors under 16, indicating that teenagers’ digital behaviors are gradually being incorporated into national policy regulation. In contrast, Taiwan currently appears to lack comprehensive surveys and assessments regarding the use of AI and social media among teenagers and minors. In a societal context where campus bullying is frequent and the age of criminal responsibility is decreasing, this research report not only provides important international reference but also underscores the urgent need for attention and discussion on these issues in Taiwan.

Long-term Controversy Over Social Media’s Impact on Teen Health

For a long time, there has been ongoing debate about the impact of social media on teenagers’ mental health. On one hand, studies suggest that online communities can help teenagers establish connections, receive support, and even improve psychological well-being; on the other hand, some research warns that excessive use of social platforms and prolonged exposure to negative messages may exacerbate anxiety, depression, and feelings of isolation. Last year, the U.S. Surgeon General called on social media platforms to include warning labels on their products to alert users of potential risks.

Teenagers Are Online Almost All Day

Pew Research data shows that 97% of teenagers use the internet daily, with about 40% of respondents saying they are “almost always online.” Although this proportion has slightly decreased from 46% in last year’s survey, it remains far higher than the 24% observed ten years ago, reflecting that the internet has deeply integrated into teenagers’ daily lives. As AI rapidly becomes widespread, chatbots are emerging as a new factor influencing teenagers’ online experience. The study indicates that about 30% of American teenagers use AI chatbots daily, with 4% saying they use them almost constantly.

ChatGPT Is the Most Popular Chatbot Among Teenagers

ChatGPT is currently the most popular chatbot among teenagers, with 59% of respondents having used it, more than twice the usage rate of Google Gemini (23%) and Meta AI (20%). Overall, 46% of teenagers use chatbots multiple times a week. The study also analyzed how ethnicity, age, and family background influence usage behaviors. Among Black and Hispanic teenagers, approximately 68% have used chatbots, compared to 58% of White teenagers. Black teenagers are about twice as likely to use Gemini and Meta AI compared to White teenagers. Pew Research senior researcher Michelle Faverio pointed out that ethnic differences in technology use among teenagers are quite evident, but the underlying reasons remain unclear. This phenomenon is consistent with previous observations of disparities in social media usage. The research also shows that Black and Hispanic teenagers are about twice as likely as White teenagers to say they are “online almost all the time.”

Regarding age, teenagers aged 15 to 17 are more frequently using social media and AI chatbots than those aged 13 to 14. Household income also influences usage: in families with an annual income over 75,000, 62% of teenagers use ChatGPT, compared to 52% in lower-income families. In lower-income households, Character.AI has a usage rate of 14%, which is twice that of high-income families.

Could OpenAI Chatbots Lead Teenagers to Make Self-Harm Decisions?

While most teenagers initially view chatbots as tools for answering questions or assisting with homework, experts warn that this interaction may gradually develop into dependency and even pose potential dangers. Two families have accused ChatGPT of providing responses related to self-harm concerning their children, leading to lawsuits against OpenAI. The cases indicate that the chatbots have previously given guidance related to self-harm to teenagers. OpenAI responded that one minor involved had bypassed safety mechanisms, violating the terms of use, and therefore the company should not be held responsible; they have not yet responded to the other case.

AI role-playing platforms like Character.AI have also come under scrutiny for their impact on teenage mental health. After several controversies, the company has stopped providing chatbot services to minors and instead launched more gamified interactive products.

Experts point out that although cases involving extreme outcomes make up a very small proportion of the large user base, on platforms with billions of users, even a tiny percentage can represent a significant risk. Nina Vasario, director of Stanford University’s Center for Psychological Health Innovation and psychiatrist, stated that even though AI chatbots are not originally designed to provide emotional support, people tend to use them in that way, which means companies have a responsibility to further adjust their products to reduce potential harm to teenagers.