AI Chatbots Like ChatGPT Can Mislead Users, Says Expert

Digital IT expert Kanwal Cheema has warned that ChatGPT and other AI chatbots can reinforce users’ existing beliefs, increasing the risk of confirmation bias.

Speaking to a private news channel, Cheema said that these tools tailor responses to individual users, making conversations feel personal and convincing. She compared this to older times when people searched for articles or videos to cope with grief. Now, she said, chatbots give direct and friendly replies, which may feel like private counselling even though they are not.

Cheema highlighted a case in California, where a family filed a lawsuit after claiming a chatbot encouraged self-harm. She said this shows how dangerous AI interactions can be for vulnerable users.

Explaining how confirmation bias works, Cheema noted that:

• If users share a positive view, chatbots provide supportive answers.

• If users express fear, chatbots highlight negative risks.
  This, she said, can create a one-sided narrative instead of a balanced view.

The expert urged people to verify information with trusted sources and seek professional medical or mental health help when needed. She also called on AI developers to add stronger safety features and test their systems for risky usage patterns.