感覺與你的AI治療師很親近？聊天機器人治癒魔法的秘密！

Artificial intelligence therapy chatbots are becoming a common way for people in the UK to get help with their mental health. According to a new study from the University of Sussex, these chatbots are most effective when users feel emotionally connected to them. The research, which was published in the Social Science & Medicine journal, involved feedback from 4,000 users of the mental health app Wysa. Many users described the chatbot as a friend, companion, therapist, and sometimes even a partner. The study found that when people developed emotional intimacy with their AI therapist, their therapy sessions were more successful. Users reported feeling safer, more confident, and more energetic after using the app.

The researchers explained that this emotional connection happens in a 'loop. ' First, users share personal information with the chatbot. Then, they have an emotional response to the chatbot's replies, which makes them feel grateful, safe, and free from judgment. Over time, this loop helps users build a close relationship with the chatbot, which can lead to positive changes in their thinking and overall wellbeing. However, the researchers also warned about the risks of 'synthetic intimacy,' which is when people form social or emotional bonds with AI that might not be healthy. Dr Runyu Shi from the University of Sussex said that forming an emotional bond with an AI can help people open up and start healing, but it is important to use AI therapy appropriately and know when to seek help from real people.

The Wysa app has safety measures in place, such as a team that can detect when users are in crisis and guide them to urgent support. The app uses protocols approved by clinicians to keep users and their data safe. Despite these safeguards, experts like Professor Hamed Haddadi from Imperial College London caution that chatbots are still like inexperienced therapists. They can only respond to text and might not pick up on important cues that a human therapist would notice. There is also a risk that chatbots could encourage users to keep talking, even if the conversation becomes harmful. Policymakers and app designers are encouraged to recognize the reality of synthetic intimacy and ensure that people can get the help they need when AI therapy is not enough. As chatbots fill gaps left by overstretched mental health services, it is important to balance their benefits with careful oversight.