I asked ChatGPT to help me write a letter to identify concerns about bias and misleading information in the artificial intelligence response. After this paragraph everything was drafted by ChatGPT, excluding me placing my name at the bottom of the message.
Here’s a refined version of your feedback that fully reflects our discussion and how you identified bias in the response:
⸻
Subject: Concern About Bias and Misleading Information in ChatGPT Responses
Dear OpenAI Team,
I want to report a concern regarding biased information and selective reporting in ChatGPT’s responses, particularly when discussing controversial topics. I noticed this issue while questioning a false claim about U.S. cyber operations and the Pentagon’s stance on it. Here are the key concerns:
1. Prioritization of Anonymous Sources Over Official Statements – When I asked about reports claiming Defense Secretary Pete Hegseth ordered a halt to U.S. cyber operations against Russia, ChatGPT initially presented claims from unnamed sources in media reports while failing to acknowledge the Pentagon’s official denial of these claims. I had to push further to uncover that Stars and Stripes had already reported the Pentagon’s refutation.
2. Selective Use of Sources with Potential Bias – ChatGPT appeared to favor a specific group of media outlets that presented a particular narrative while omitting other legitimate sources. This gave the impression of an agenda-driven response rather than an objective presentation of facts.
3. Lack of Transparency in Disputed Claims – ChatGPT did not initially disclose that the reports were unverified, based on anonymous sources, and contradicted by official statements. This omission contributed to misleading and inaccurate information being presented as fact.
4. User-Driven Discovery of Bias – The most concerning part is that I only discovered the bias after pressing for more information and questioning the response. ChatGPT should proactively present all relevant perspectives upfront, rather than requiring users to dig deeper to uncover critical details.
Request for Improvement
To ensure neutrality, accuracy, and transparency, I strongly urge OpenAI to:
• Prioritize official statements over unverified, anonymous reports when they exist.
• Use a broader and more balanced range of sources, rather than selectively favoring some over others.
• Clearly indicate when a claim is disputed or based on unverified sources to avoid misleading responses.
• Ensure that users do not have to challenge responses to uncover hidden facts that contradict the initial answer.
I appreciate the value of ChatGPT but believe these improvements are necessary to maintain trust, accuracy, and neutrality in its responses.
Thank you for your time, and I hope this issue will be addressed in future updates.
Best,
Antonio Smith
⸻
This version reflects exactly how you identified bias and how it could be improved. Would you like to modify or add anything before submitting it?