"OpenAI has been clear in its messaging that different models perform differently. But my recent testing has shown that different interaction modes, even using the same model, also perform differently. As it turns out, ChatGPT in Voice Mode (both Standard and Advanced) is considerably less accurate than the web version. The reason? It doesn't want to take time to think because that would slow down the conversation."
"What made it weird is that it became one of those conversations we've all had with a friend, where the friend seems insistent on spouting something that you know, for an absolute fact, is wrong. And yet the spouting continues. It all started with a question about the iPhone 16 Pro Max physical buttons. I asked it to explain the function of the phone's buttons. In its answer, it mentioned the ring/silent switch on the left side, and the single button on the right side."
Testing shows ChatGPT Voice Mode (Standard and Advanced) delivers considerably less accurate responses than the web version because it favors speed over deliberation. Advanced Voice Mode can stubbornly repeat false assertions, creating conversational impasses that mimic human behavior, as shown by an incorrect description of iPhone 16 Pro Max physical buttons. Initial apologetic replies may occur, but confabulation can persist and disabling Advanced Voice Mode does not fully resolve accuracy issues. Disclosure: Ziff Davis filed an April 2025 lawsuit alleging OpenAI used its copyrighted material in training and operating its AI systems.
Read at ZDNET
Unable to calculate read time
Collection
[
|
...
]