Q. Is AI biased?

The short answer is: yes. Generative AI is biased, because the data it uses it biased.

Generative AI tools (such as ChatGPT) use large amounts of data from the internet in order to train their models. That data is more likely to be in English and to represent certain viewpoints. While different generative AI tools have tried to implement guardrails to prevent biases, these guardrails aren't perfect. You can expect that some of the text and other content you generate from AI tools to have gender, racial, and other forms of biases.

Chatbots also reinforce biases. If you ask Chat GPT or Copilot to provide evidence for something, it generally will. You can utilize prompt engineering to try and ensure that your answer is as unbiased as it can be. Instead of asking a chatbot to tell 3 reasons why screentime is bad for children, ask for benefits and concerns of letting children have screentime.