December 27, 2025

Can A.I. Be Trusted?

DISHONESTY AND ARTIFICIAL INTELLIGENCE

On December 27, 2025, I asked AI a couple of questions:

• Does A.I. give different answers to the exact same question but from different people?
• If A.I. gives different answers to the exact same question but from different people, then can A.I. be trusted?

The answers I got were quite interesting…

My first question:  Does A.I. give different answers to the exact same question but from different people?

A.I.’s response:  Yes, AI often gives different answers to the same question, even from different people or the same person at different times, due to built-in randomness (stochasticity), busy server loads, different model versions, varying contexts, and unique prompts, making each response a fresh generation rather than a database retrieval. This variability stems from AI’s probabilistic nature, aiming for creative, human-like output, but can be controlled with settings like “temperature” for more consistent results

My second question:  If A.I. gives different answers to the exact same question but from different people, then can A.I. be trusted?

A.I.’s response:  No, an AI’s differing answers to the same question mean you can’t fully trust it without verification, as variations stem from its probabilistic nature, system load, and training data, making it a creative tool needing critical cross-checking with reliable sources, not an infallible oracle. Treat it as a starting point, not the final word, by asking multiple times, checking facts, and being skeptical of confident-sounding answers, especially for crucial information, because it generates patterns, not absolute truths.