By skyforbes ChatGPT tends to try to be nuanced and shares both the pros and cons with things. It may say a historical event or a phenomenon may have both good and bad sides with it, while Geminini is more likely to say something is either good or bad having a more black/white view.
Even on mundane things they are likely to reply differently. ChatGPT is from my experience more likely to claim illfitting clothing are uncomfortable, while Geminini is more likely to use the word "pain" to describe something.
When it comes to extreme situations like countries with lack of medical resources, Geminini is more likely to say long term intubation without anesthetics/painkillers/sedation is common while ChatGPT is more likely to say it's not and that it's more common for patients to refuse treatments if no medicines is available even if their lives won't be saved. Geminini is more likely to claim painful surgeries done without consent and anesthetic is done in poorer countries, while ChatGPT says it's not true for adults.
AI general speaking isn't reliable, but it's fascinating how different they reply.