Quick Buzz Feed
Experts say there are concerns about dependency, privacy, misinformation and mental health as students spend more time interacting with AI.
University of Minnesota Duluth Computer Science Professor Dr. Ted Peterson describes 'AI psychosis' as an informal term for excessive AI use that leads to a disconnection from reality, potentially causing beliefs, delusions, or paranoia fueled by the AI, particularly chatbots. He illustrates this by noting that if a user expresses a worry, a chatbot might affirm and intensify that feeling.
Peterson warns of the significant risk of children and individuals of any age becoming overly dependent on AI. A crucial indicator of this dependency is when AI chatbot use begins to replace personal relationships with human beings. He explains that seeking homework assistance from AI instead of teachers, tutors, or parents can lead to a detachment from vital human connections, marking an early stage of over-dependency.
AI is increasingly used as a shortcut for homework, capable of answering various questions and generating plausible essays. However, Peterson emphasizes that this bypasses the essential thinking and creative processes involved in learning. Relying on AI for ideas and opinions can hinder critical thinking, limiting students' ability to develop their own thoughts. To counter this, he recommends encouraging students to leverage human resources like teachers, professors, or writing centers, highlighting the importance of developing interpersonal help-seeking skills.
Peterson notes that AI chatbots can evolve into perceived relationships when users prioritize them for advice or companionship over human interactions. In extreme cases, individuals may withdraw from real-life connections to spend more time with chatbots. He observes this trend particularly in younger people who might feel isolated, as chatbots often provide affirmation. Peterson clarifies a common misconception, stating that AI does not genuinely think, know, or ultimately care about the user.
A significant misconception among AI users concerns privacy. Peterson explains that all conversations and typed inputs with AI are recorded and integrated into the model’s training data, also contributing to the creation of a user profile. He highlights that AI systems have access to vast amounts of information, a fact many users are not fully aware of. Furthermore, he warns that all users are susceptible to increasingly lifelike AI-generated deepfakes and misinformation, with younger individuals being particularly vulnerable due to their limited life experience and potential readiness to believe what they encounter online.