Unmaking Sense

Episode 8.15: That chatGPT “fabricates facts and figures” only makes it more remarkable, more human.

Inference, interpolation and educated guesswork are characteristics of intelligence, not black marks to be held against it. Will Knight’s critique of chatGPT in “Wired” serves to enhance its status, not diminish it.

Episode 8.14: ChatGPT has faults; the remarkable thing is that they are exactly like ours.

Numerous commentators have been queueing up to pour tepid water over the achievements of the OpenAI conversational chat bot chatGPT. What is of particular interest about their dismissive or questioning comments is that they accuse the chatbot of having as significant faults precisely the kind of things of which we would accuse one another. For example, arguing that the chat bot “fabricates facts and figures” because of the way it operates would seem to ignore the fact that human beings do just the same. That we do not understand how the chatbot draws its inferences, and perhaps accuse it unfairly of using statistical patterns, rather than “connecting words with meanings” seems to suggest that we understand how humans understand language when in fact we don’t. You and I have a huge number of instances of the occurrence of every word in our native tongue stored in our heads, and somehow we make sense of them, but nobody knows how. Somehow we connect the word table with what the word table means. This is a problem that goes back to Plato. Who is to say how we do that? Nobody has any idea, and that we perform some kind of statistical matching process may well be as good an explanatory candidate as any. So it’s unfair to accuse chat GPT of failings that we have ourselves. That just means that it is closer to joining the human conversation as a fully-fledged participant than we might otherwise have thought.

Episode 8.13: As the embodiment of all our values and knowledge, is chatGPT like some modern deity?

We consider social constructivism as the academic discipline that studies the notion that in our religions, and in particular in our deities, we find the result of the projection, or the apotheosis of social values in some universal absolute binding supreme power. If that thesis is accepted, and for many people it is not, could it not be the case that something like chatGPT as the embodiment of all our social values and knowledge, a suitably advanced artificial intelligence, could be regarded as some kind of modern deity? (Poor quality recording.)

Episode 8.12: How can education based on self-interest give a knowledge-economy the skills it needs?

Allowing children to dictate the trajectory of their own learning based on interests will not - or so says the educational establishment - provide a knowledge economy with the skilled workers it needs. So using chatGPT to facilitate such an education is a bad idea. We reject this objection as self-serving and short-sighted.

Episode 8.11: How can we possibly allow children to dictate their education based on their interest?

The second of three predictable objections: that children don’t know what they don’t know. Answers include the personalised curriculum, personalised assessments and personalised accreditation.

Episode 8.10: How can we safeguard against the abuse of chatbots like chatGPT to control thinking?

When asked how Adolf Hitler would use AI, chatGPT answered “to manufacture better propaganda”. We consider the chilling implications.

Episode 8.09: How chatGPT facilitates personalised learning, assessment and accreditation.

The educational establishment will try to deny that personalised learning, assessment and accreditation are possible because they sound the death-knell of a multi-trillion dollar industry. But they will be good for humanity, and good for every individual enabled to define their own educational trajectory in pursuit of their own unique life without reference to an educational behemoth that tries to dictate what we should learn, at what level of difficulty and depth, in what order and when. We each become our own AlphaZero, effectively self-educating, learning what interests us at our own pace in our own time to whatever level of depth and difficulty we can manage.

Episode 8.08: How chatGPT heralds an unprecedented revolution in personalised education.

Since the dawn of human history, we have been forced to do education through older human beings teaching younger human beings what they know, often in institutional centres of learning. Until the invention of printing, access to resources was severely restricted both because of their scarcity, and because of their value, and only with the advent of the computer and the Internet has that really changed m. We no longer needed access to physical libraries then, and now we no longer need human experts because chat bots will be able to provide better information more quickly and a way that is far more tailored to the particular interests and abilities of each individual pupil. Assessment can similarly be personalised or even become unnecessary. The world of education has suddenly changed.

Episode 8.07: Why an AI like chatGPT makes mistakes. Some of the less obvious mistakes explained.

Further detail on why chatGPT and any other neural net AI will almost certainly make mistakes. Training data, overfitting, interpolating, inferring and generating errors by inference. Why increasing the complexity of a model to avoid such problems may allow the neural net to overfit to include noise in the training data, and so introduce another source of potential error.

Episode 8.06: The Singularity and the significance of OpenAI’s chatGPT in the debate about it.

We discuss three kinds of singularity before turning to “The Singularity” in terms of superhuman intelligence and solving intractable problems.