I am not using AI for the purposes the most of you do.

I am neither a programmer nor a developer.

But I feel the "big players" on this field should be honest about their intentions.

I tried to find, who is the owner of the photo. Who shot it.

👉 Chatgpt - 4o: Gave me a name, which turned out to be false, since there was no evidence. It admitted that after a follow up question.

👉 Gemini 2.5 Pro: Gave me another false name. Admitted that it was wrong afterwards as well.

👉 Gemini 2.5 Flash: Admitted it didn't know who is - that's strange, since the more developed model gave me false information.

👉 Perplexity Pro: Admitted it didn't know.

👉 Grok 3: Admitted it didn't know.

➡️ It's disappointing, especially from OpenAI and Google. No costumer expects to take a 100% accurate answer about everything, but at least they shouldn't mislead. This is common logic.

Why the developers of those tools, with all their "wisdom" cannot predict that?