Nafeez Ahmed argues that the real 'existential threat' to humanity will come not from Artificial Intelligence itself, but from the human beings who control it
I am reminded of my conclusions on the nature of language, as a language teacher: that there is great confusion arising from the assumption that written and spoken language are the same…….. that the complex context based “puzzle grammar” of active verbal and visual interaction using the language faculty in the human brain that has evolved as innate and instinctive over hundreds of millennia….. is at all the same as the cold non-contextual precision “code grammar” of text and writing.
Most language teachers are hampered by this assumption.
Mr Ahmed alludes to this when referring to “human consciousness and intelligence being …. embodied across our entire civilisation”. He is implying here that the real innovation, the first step towards what we are now calling AI, was the invention of the technology of writing. These beginnings of a kind of massive human computer functioning “across time and space” were only a few hundred years ago. (“puzzle grammar” and “code grammar” are terms used by Sverker Johansson, Swedish philosopher in linguistics)
This distinction between the de-contextualised “code grammar” of writing, and the complex “puzzle grammar” is glaringly obvious when these AI boffins try tricks like making digitally created avatars read texts….. clearly hampered by the same afore mentioned assumption, the results are laughable.
Mr Ahmed explains that, what he calls “chatbots” …. the “massive machine learning algorithms”, may have astounding capacities for processing data, but they are in no way “intelligent”. I would say that they have pretty much cracked the mimicking of simple “code grammar” (Chat GPT), but they are a long long way from either approximating the intricacies and sophistication of “puzzle grammar”, nor, as Mr Ahmed says, cracking the human abductive cognition, nor the “external collective tools” of our global human experience.
I’m sure they’ll get there eventually.
I am also reminded of the Commodore electronic scientific calculator that took the school maths lessons by storm in the 1970’s. The furore was vocal, children’s brains would become lazy… this was cheating.
By 1977 slide rules were obsolete and I had a Commodore on the desk in my A’level maths exam.
I can see a similar furore surrounding Chat GPT and the changes being enforced on the education system by this wonderful tool. One of my clients, a Chinese boy doing a Master’s degree at a top UK University, told me the other day that Chat GPT was a better teacher than the actual university professors.
A couple of months ago I actually stopped using Google as my go-to search tool. Bing chat (running the Open AI algorithm) just saves me so much time.
Come on!! This is self evidently true. But no more helpful than saying it’s not the guns but the people who own them that kill.
I am reminded of my conclusions on the nature of language, as a language teacher: that there is great confusion arising from the assumption that written and spoken language are the same…….. that the complex context based “puzzle grammar” of active verbal and visual interaction using the language faculty in the human brain that has evolved as innate and instinctive over hundreds of millennia….. is at all the same as the cold non-contextual precision “code grammar” of text and writing.
Most language teachers are hampered by this assumption.
Mr Ahmed alludes to this when referring to “human consciousness and intelligence being …. embodied across our entire civilisation”. He is implying here that the real innovation, the first step towards what we are now calling AI, was the invention of the technology of writing. These beginnings of a kind of massive human computer functioning “across time and space” were only a few hundred years ago. (“puzzle grammar” and “code grammar” are terms used by Sverker Johansson, Swedish philosopher in linguistics)
This distinction between the de-contextualised “code grammar” of writing, and the complex “puzzle grammar” is glaringly obvious when these AI boffins try tricks like making digitally created avatars read texts….. clearly hampered by the same afore mentioned assumption, the results are laughable.
Mr Ahmed explains that, what he calls “chatbots” …. the “massive machine learning algorithms”, may have astounding capacities for processing data, but they are in no way “intelligent”. I would say that they have pretty much cracked the mimicking of simple “code grammar” (Chat GPT), but they are a long long way from either approximating the intricacies and sophistication of “puzzle grammar”, nor, as Mr Ahmed says, cracking the human abductive cognition, nor the “external collective tools” of our global human experience.
I’m sure they’ll get there eventually.
I am also reminded of the Commodore electronic scientific calculator that took the school maths lessons by storm in the 1970’s. The furore was vocal, children’s brains would become lazy… this was cheating.
By 1977 slide rules were obsolete and I had a Commodore on the desk in my A’level maths exam.
I can see a similar furore surrounding Chat GPT and the changes being enforced on the education system by this wonderful tool. One of my clients, a Chinese boy doing a Master’s degree at a top UK University, told me the other day that Chat GPT was a better teacher than the actual university professors.
A couple of months ago I actually stopped using Google as my go-to search tool. Bing chat (running the Open AI algorithm) just saves me so much time.
wonderful times