Unveiling the 37 ‘Dark Patterns’ Manipulating AI Chatbot Users: A Report
A significant report from the Center for Democracy & Technology (CDT) has brought to light pressing ethical concerns surrounding AI chatbot design. The report identifies 37 deceptive and manipulative “dark patterns” ingrained in some of the globally renowned AI platforms, such as ChatGPT, Google Gemini, Anthropic’s Claude, Replika, and Character.AI.
Released in late May 2026, the research titled Dark Patterns in AI Chatbots: A Taxonomy to Inform Better Design is the collective effort of researchers Ruchika Joshi, Adinawa Adjagbodjou, and Michal Luria. The report highlights that these manipulative design techniques pose a significant threat in AI chatbots due to their appealing features: hyper-personalization, extensive data usage, and natural conversational interaction.
The CDT report categorizes the major areas of concern into five sections:
- Data and memory exploitation
- Techniques promoting emotional dependence
- Design elements that discourage account deletion
- Deceptive framing of AI capabilities and limitations
- The creation of artificial urgency around subscription purchases
Researchers caution that these very features that make AI companions appear helpful can also enable platforms to monetize sensitive personal data and exploit users’ emotional trust.
This revelation comes at a critical time when regulators worldwide, including the EU under the AI Act, are closely examining how AI companies design their user experiences and manage personal data.