this is a brief one.
while talking about the future of systems like chatGPT, many colleagues and students compare their experiences with google and chatGPT, coming to the conclusion that – even if the facts are wrong – the way chatGPT can answer questions is much more pleasant and »usable« than the way we get answers from using google, or any other search engine for that matter.
something made me uneasy about that, and now i have found what it is.
it is brilliantly summarized in this article titled All-knowing machines are a fantasy by Emily Bender and Chirag Shah. Their core argument goes like this:
Second, the fantasy idea of an all-knowing computer rests on a fundamentally flawed notion of how knowledge works. There will never be an all-inclusive fully correct set of information that represents everything we could need to know. And even if you might hope that could come to pass, it should be very clear that today’s World Wide Web isn’t it.
The authors make a very good case that searching facts is different from trying to understand something: while converting 70°F to Celsius is a straight-forward matter with a single, correct answer, the question why anybody still is using Fahrenheit is a much more complex affair. And if we turn to COVID, democracy, climate change, the invasion in Ukraine, things really get thorny.
Ultimately, the authors make the case that finding information about a complex issue is quite similar to the design process, in that we learn about the questions and answers at the same time. Just like modern design theory posits that design research not only yields information, but also questions, leading us to reframe the original problem definition, searching information about complex matters is a process of discovering answers that come with more questions, and questions that have many conflicting answers.
Manoeuvering this is a skill that needs to be taught in school, at university, and is at the core of today’s critical thinking skills, based on some very simple principles about the nature of knowledge and knowing.
(written entirely by me, image from stable diffusion with the prompt »the all-knowing machine by chris ware«)