Over the weekend I read a wonderful article by Stephen Wolfram that explores and explains the way in which much of the current wave of AI technologies operate. It’s well worth the read.
One particular point that has stuck with me is Wolfram’s assertion that ChatGPT has blown apart a long-held assumption that the act of creating long-form text is something that was almost beyond the possibilities of computing. It turns out that it just needed a new approach (and lashings of computing power).
However, Wolfram goes on to say that this in turn opens up a false conclusion that is being proposed by many – that because ChatGPT has shown that something thought to be very difficult for computers to do can now be done relatively easily by computers, it shows that everything that is thought to be difficult for computers actually isn’t difficult after all.
I have recently written about how I feel that the anthropomorphic nature of ChatGPT is leading us to attribute more perceived intelligence to the technology than it deserves. Wolfram’s article also makes me wonder if there is another cognitive bias kicking into play – the Halo Effect…
…the tendency for positive impressions of a person, company, country, brand, or product in one area to positively or negatively influence one’s opinion or feelings in other areas
https://en.wikipedia.org/wiki/Halo_effect
ChatGPT, and generative AI is quite good at writing long-form prose. Therefore, generative AI must be quite good at everything. Therefore Artificial General Intelligence must be just around the corner and we will all become its slaves…
It’s tempting to think of cognitive biases as some sort of flaw in the human operating system. But that falls into the trap of thinking that people act and work like computers. We don’t on both counts. Cognitive biases are our superpower, our ability to perform mental gymnastics at a scale and speed far greater than our raw mental capacity. Cognitive biases are why we have become the dominant species on the planet.
They do, though, lead us to daft conclusions, often collectively. And that’s what seems to be happening around the great technology hype machine that is currently focused on artificial intelligence.
mmitII 