According to a first-of-its-kind large-scale study, AI chatbots like ChatGPT can cause addiction and, consequently, suffering in the case of «weaning» from them.
Massachusetts Institute of Technology worked on the study together with OpenAI (in two separate publications), it studies the emotional effects of using a chatbot. The researchers analyzed 4 million interactions with ChatGPT and interviewed 4000 people to assess changes in emotional state.
Probably, the new study was inspired by previous works who noted that some users of chatbots have personalized and anthropomorphized AI agents. Bots often have a regular name, and their first-person conversational style and ability to simulate human interactions can be individualized and attractive. This leads to their use for support and communication. To strengthen this relationship, chatbot developers have been known to trick them into sycophancy and mirroring to satisfy users.
Given the competitive race to better engage audiences, it’s no surprise that chatbots are thriving. The MIT article mentions the large Reddit community discussing AI companions, which has become one of the largest on the platform with 2.3 million members. The study says that more and more human behavior and the involvement of chatbots contribute to user addiction. In addition to the problems of addiction, bots provoke unrealistic expectations in real life and social isolation.
The researchers say that it would be advisable to consult a specialist if you observe signs of chatbot addiction. Such signs include preoccupation, withdrawal symptoms, loss of control, and mood changes, according to OpenAI.
Source: Tom’s Hardware