ChatGPT: never underestimate the power of the spoken word
OpenAI has updated the paid version of ChatGPT on smartphones with a voice generator able to receive spoken questions and answer them with five different voice options, potentially sidelining Alexa, Siri or Google Assistant. In addition, it can accept images as activation commands and comment on “what it sees”.
This is stealing a march on other generative algorithms, further stimulating competition in an already dynamic environment, making AI even more accessible, and opening the door to a whole new world of possibilities for interaction with these powerful assistants.
However, it is also important to understand the big differences between text and voice interfaces: text interfaces tend to involve a more reflexive processing of the information received — due to the need toread and interpret the text, if not, as in some specific assistants, to review the sources used to prepare the answer. In the case of “spoken” answers, we tend to process the information instantaneously, and are less likely to question it.
This brings me back to the article I wrote in January on the impact of generative algorithms on critical thinking, “ChatGPT and the decline of critical thinking”: we used to be worried about “outsourcing” our critical thinking and judgment to an algorithmic assistant, but we now have…
