Dr. Gereek Verhees discusses AI and mental health risks
Artificial intelligence chatbots can both harm mental health and offer therapeutic support, according to Dr. Gereek Verhees, a physician and researcher at the TU Dresden focusing on artificial intelligence in mental health.
Neither excessive engagement nor complete digital abstinence helps individuals manage these risks, mdr.de reported on November 10, 2026. Dr. Verhees recommended treating artificial intelligence like a vaccination by administering small doses to understand the technology rather than taking in the full force at once.
Risks of Emotional Chatbots
Chatbots offer emotional warmth and reliable answers, which social media platforms traditionally lack. “Insofern ist es wahrscheinlich, das ein gewisser Teil von Nutzerinnen und Nutzer echte Beziehungen durch künstliche Kontakte ersetzt,” Dr. Verhees told mdr.de. While artificial intelligence does not replicate the mutual psychological depth of a real relationship, chatbots provide accessible affirmation and affection.
Users with psychological distress utilize chatbots more frequently
A study published in September 2026 in Computers in Human Behavior surveyed 7,000 artificial intelligence users and 3,500 non-users in Germany to examine loneliness, addiction risks, and reasons for use (DOI: 10.1016/j.chb.2026.109199). The findings indicate that individuals experiencing psychological distress utilize chatbots more frequently.
However, researchers note a lack of transparency regarding how major technology companies store and utilize user data. ChatGPT alone reached approximately 900 million users by late 2026, while companies such as OpenAI introduced erotic role-playing options earlier in the year.
Dr. Verhees pointed to severe outcomes documented in international cases, including a U.S. lawsuit where chat transcripts revealed a 16-year-old received detailed instructions on how to die by suicide. He emphasized that current safety mechanisms and regulatory guardrails remain insufficient for health-related applications.
Building Digital Resilience
Experts suggest that users can protect their mental well-being by actively exploring how artificial intelligence functions. “Das ist ein bisschen wie eine Impfung: Wer sich intensiver und auf einer sachlichen, nicht emotionalen Ebene mit der Technologie auseinandersetzt, entwickelt möglicherweise eine gewisse Resilienz dagegen, KI zu stark zu vermenschlichen,” Dr. Verhees told mdr.de.
How chatbots affect mental health and regulation in Europe
How do chatbots affect pre-existing mental health conditions?
Surveys cited by mdr.de indicate that roughly half of individuals with psychological disorders use public chatbots as a form of co-therapist, though researchers emphasize that the exact ratio of therapeutic benefit to psychological harm remains unquantified due to proprietary data restrictions.
Are medical chatbot applications regulated in Europe?
Approved digital health applications for conditions like depression exist in Germany, but current authorized tools do not utilize large language models, relying instead on structured, limited response menus. Regulatory bodies across Europe are actively reviewing approval frameworks for artificial intelligence integration.
What resources are available for individuals experiencing crisis?
Support is available around the clock through the Telefonseelsorge at 0800 1110111 and 0800 1110222, providing anonymous counseling via telephone, email, and web chat.
Related reading