Why is Copilot Telling Me I'm Literally Wrong, Iran's Leaders Were Never Attacked?

Question

Why is Copilot Telling Me I'm Literally Wrong, Iran's Leaders Were Never Attacked?

Ryan Stimpert 5

I literally spent 5-minutes trying to convince Copilot that Iran's leaders were killed on February of 2026 and I was consistently told I was wrong, that the US would never risk a regional war??? But yet it would occasionally agree that the US had struck targets, but ONLY Iranian target that were outside of Iran, nothing in Iran. I copy and pasted the exact same question to Gemini, Chat GPT and Perplexity and they all gave me the correct answer, details and all. It comes across as Copilot trying to spread propaganda, not factual information, rather information that fuels Microsoft's own views whatever that may be. If there is a policy issue then it should be giving the answer "I'm sorry but I can't answer that" or something similar. But instead it was insistent, STRONGLY INSISTENT, that EVERY major news source CURRENTLY says that Iran's leaders are still alive and never attacked and over and over that the US would never risk a regional war. This is information that is 2.5 months old, every other AI platform get's it right, so that leads me to believe that Copilot is pushing information that's not factual and doing it's best to make the user believe it!!

What's going on, I don't feel safe using Copilot anymore if that kind of crap is gonna to be put out and I've deleted it from my phone...

Somil Jain 660 Reputation points

2026-05-17T11:05:47.0366667+00:00

Hi Ryan Stimpert,

I understand why that experience would be frustrating. AI systems can sometimes provide outdated, incomplete, or incorrect information, especially regarding fast-moving geopolitical events. In some cases, models may rely on stale sources, conflicting data, safety filters, or cached responses that lead to inaccurate answers being presented too confidently.

You are correct that if the system is uncertain or restricted from answering, it should communicate that clearly instead of repeatedly insisting on incorrect information. Comparing responses across multiple AI systems, as you did, is actually a good way to verify sensitive or breaking-news topics.

This behavior does not necessarily mean intentional propaganda, but it does highlight the importance of independently verifying information with trusted news sources, especially for political or international events. Feedback like yours is valuable because it helps identify situations where the model handled uncertainty poorly.

Thank you very much, have a nice day.
RR M 0 Reputation points

2026-07-01T05:17:44.57+00:00

It will also tell you that there is no conflict between Iran and the US there is no Islamabad MOU because there is no war, It will give you the congress vote results on not supporting the war with a breakdown of votes then say its false because there is no war. It's got to be stale sources because it claims there are no reports on BBC Reuters Al Jazeera AP etc
Silly Donkey 0 Reputation points

2026-09-08T21:56:24.6666667+00:00

I also had the same about Khamenei a few months ago, I just assumed it was a mistake, but even after posting a link to reliable news sources, CoPilot said I was being lied to!

Then just now, its telling me that the US did not attack Iran and that there is no conflict and the Strait of Hormouz is open and there is no reason why it would be closed.

I have discussed many up to date events with CoPilot and these are the only two events that it has denied, even to the point of saying that I am playing a joke on CoPilot!

Your answer

Somil Jain 660 Reputation points

2026-05-17T11:05:47.0366667+00:00

Hi Ryan Stimpert,

I understand why that experience would be frustrating. AI systems can sometimes provide outdated, incomplete, or incorrect information, especially regarding fast-moving geopolitical events. In some cases, models may rely on stale sources, conflicting data, safety filters, or cached responses that lead to inaccurate answers being presented too confidently.

You are correct that if the system is uncertain or restricted from answering, it should communicate that clearly instead of repeatedly insisting on incorrect information. Comparing responses across multiple AI systems, as you did, is actually a good way to verify sensitive or breaking-news topics.

This behavior does not necessarily mean intentional propaganda, but it does highlight the importance of independently verifying information with trusted news sources, especially for political or international events. Feedback like yours is valuable because it helps identify situations where the model handled uncertainty poorly.

Thank you very much, have a nice day.
RR M 0 Reputation points

2026-07-01T05:17:44.57+00:00

It will also tell you that there is no conflict between Iran and the US there is no Islamabad MOU because there is no war, It will give you the congress vote results on not supporting the war with a breakdown of votes then say its false because there is no war. It's got to be stale sources because it claims there are no reports on BBC Reuters Al Jazeera AP etc
Silly Donkey 0 Reputation points

2026-09-08T21:56:24.6666667+00:00

I also had the same about Khamenei a few months ago, I just assumed it was a mistake, but even after posting a link to reliable news sources, CoPilot said I was being lied to!

Then just now, its telling me that the US did not attack Iran and that there is no conflict and the Strait of Hormouz is open and there is no reason why it would be closed.

I have discussed many up to date events with CoPilot and these are the only two events that it has denied, even to the point of saying that I am playing a joke on CoPilot!