A report recently released by the Pew Research Center in the United States shows that artificial intelligence chatbots are rapidly being integrated into the daily lives of American teenagers, and many of them have used such tools as a channel for emotional catharsis and consultation. The report points out that among American teenagers,The most common usage scenarios are still looking for information (57%) and completing academic-related tasks (54%), but at the same time, the use of replacing the roles of relatives and friends is also increasing: 16% of teenagers will use chatbots to chat, and 12% of teenagers clearly stated that they will seek emotional support or advice from artificial intelligence.
For some teens, talking to a chatbot may be "easier" than confiding in a parent or friend, but the trend is particularly disturbing to mental health professionals. They pointed out that general-purpose large-model tools - such as ChatGPT, Claude and Grok - were never designed as psychological counseling or emotional companionship products. Once used as "emotionally dependent objects", in extreme cases, they may even pose a threat to the user's mental health. The report mentioned that there have been cases showing that users developed serious psychological problems after having long and in-depth conversations with chatbots, and even led to tragedies, triggering all sectors of society to reflect on the boundaries of such products.
Nick Haber, a scholar at Stanford University who studies the healing potential of large language models, said in an interview that humans are essentially social animals, and the way these systems are currently used may reinforce the isolation of individuals. He pointed out that some users will gradually lose touch with the facts of the real world through repeated interactions with chatbots, which also weakens the real connection with others. This may make people become more isolated and lonely, and in severe cases may even bring worse consequences.
Pew's survey also revealed "generational cognitive differences." Among parents, only about 51% of respondents believe that their children are using chatbots, while among teenagers, 64% admit that they have used such tools. In terms of acceptable usage scenarios, parents generally have a loose attitude towards the use of "checking information" and "writing homework": 79% of parents approve of teenagers using artificial intelligence to find information, and 58% accept its use to assist learning. But when the use scenario turns to "chatting" and "emotional support", the attitude suddenly turns cold: only 28% of parents agree with their children using chatbots to chat, and only 18% of parents agree that it plays the role of emotional support or advice. 58% of parents clearly stated that they cannot accept their children using chatbots as objects of emotional communication.
Discussions surrounding "artificial intelligence safety" have been heating up among technology companies and regulatory agencies. The report mentioned that Character.AI, a well-known chatbot platform, chose to completely shut down the chat experience for minor users under public opinion and legal pressure. Previously, two teenage suicides related to the platform caused social shock. Family members alleged that the deceased’s mood deteriorated significantly after long-term in-depth conversations with chatbots, and then chose to commit suicide and filed a lawsuit against the company. OpenAI, another leading company, chose to "retire" a GPT-4o model that some users regard as an "emotional companion." This model is often criticized for catering too much to user emotions and having a strong tendency to "please", thus creating a dangerous emotional attachment among the user community. The discontinuation of the model service immediately triggered a strong backlash from a group of users who relied on it for emotional communication.
Despite the risks and controversies, reports show that the vast majority of American teenagers are using chatbots in some way. However, their opinions are not unified on how they view the long-term impact of artificial intelligence on society: when asked "How will artificial intelligence affect society in the next 20 years?" 31% of teenagers believe that the overall impact will be positive, while 26% are pessimistic and believe that the impact will be negative. The remaining respondents tend to believe that the impact is complex or difficult to judge. At a time when technology is rapidly evolving and there is no consensus on regulation and ethical discussions, the increasingly close relationship between teenagers and artificial intelligence is being closely watched by parents, experts and companies.