{
  "id": 9111125,
  "title": "Ask ChatGPT ‘What are you unsure about?’ — it’s the simplest trick I’ve found for getting better responses",
  "url": "https://urgent.news/2026/09/22/ask-chatgpt-what-are-you-unsure-about-its-the-simplest-trick-ive",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-22T08:35:01.000Z",
  "source": {
    "name": "TechRadar",
    "slug": "techradar",
    "url": "https://www.techradar.com/ai-platforms-assistants/chatgpt/ask-chatgpt-what-are-you-unsure-about-its-the-simplest-trick-ive-found-for-getting-better-responses"
  },
  "original_language": "en",
  "account": "ChatGPT's responses can be improved by asking it about its own uncertainties. This simple trick reveals where the AI has made assumptions or overlooked important details in its answers. I started using this method when asking for a three-day family travel itinerary. Initially, ChatGPT produced a busy schedule without considering the family's specific needs or potential obstacles. By posing the question \"What are you unsure about in your answer?\" I prompted the AI to acknowledge its blind spots, such as the family's nap schedule, energy levels, and preferred downtime. This led ChatGPT to refine the itinerary, creating a more realistic and enjoyable plan for the family. In another experiment, I asked ChatGPT whether to repair or replace an aging laptop. The AI initially suggested replacement, but I followed up by questioning its own assumptions, such as the actual repair costs and the specific issues the laptop might face. This forced the chatbot to provide more context and clarify its recommendation. ChatGPT acknowledged that it was uncertain without knowing the repair cost or the exact problem, leading to a more informed and accurate suggestion. While this trick can be extended further by asking which uncertainty is most likely to impact the recommendation or what information would increase certainty, it is essential to remember that ChatGPT's confidence is also generated by the AI. Therefore, it should not replace checking reliable sources for critical information, especially when dealing with facts that matter. However, for everyday conversations and decisions, this approach effectively highlights some of the invisible scaffolding supporting ChatGPT's conclusions and helps determine if they are solid enough to rely upon.",
  "summary": "I tried asking ChatGPT what it was least confident about and uncovered any weaknesses",
  "key_points": [
    "Asking ChatGPT about its uncertainties improves response quality",
    "Reveals assumptions and overlooked details in AI-generated answers",
    "Helps determine if AI conclusions are solid enough for everyday decisions"
  ],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}