AI Learning: It’s a Tool, Not the Terminator

I?recently had the honour?of?speaking?on?a?webinar?hosted by Ocean Academy, a?learntech company from Ukraine, about?AI?and?learning.?It went well, but I?was struck by the way people were framing the questions they asked me. Questions like: ‘what do you make of AI-based digital humans taking jobs away from regular workers?’ and: ‘How do you imagine a collaboration between AI and humans (trainers, designers)?...?what balance should be found between them?’.

There is a lot to say about AI and learning, which I have been researching and writing about?for a good number of?years, but?faced with the assumption?underlying?questions phrased?in this way?– that?AI is some kind of alien intelligence, already on a level with?human intelligence – I think it’s worth taking a step back.

Can we really have a sensible discussion of?what?A.I.?means for?learning with?assumptions like these?as our starting point?

There is a natural human tendency to reify abstract things such as A.I., shaping a mass of zeros and ones into a form about which we can feel emotions of some kind.?To this our algorithm-driven media adds a freight of rage,?fear?and?eroticism as it does to everything it wants?to make more clickable. The result is?that?we’ve been?led to think of AI as something?similar to?ourselves, with similar properties of sentience, identity and agency?–?only much smarter and possessed of an unknowable, potentially malign intentionality. But none of this is true of A.I., at least?as we currently see it.

A.I.?has no intentionality, no desires of any kind. It can’t, in any real sense of the word, ‘know’ anything at all. According to internet pioneer Jaron Lanier writing in the New Yorker, ‘the most pragmatic position is to think of A.I. as a tool, not a creature’.?More like a torque wrench than The Terminator.

Generative AI in particular, the form of AI currently provoking most of the fear that we are imminently about to be replaced, or taken over, only appears so ‘lifelike’ in its responses because it plays back to us our own habits of speech. It has been trained on a vast subset of everything ever said or written on the internet by humans. It is, Lanier says, ‘a mashup of?us’.Perhaps the reason that?ChatGPT?has been such a massive media?event?is?that it?has a natural language interface. For the first time, we can interact with an AI using ordinary human languages, not code. A field?that?was previously all about maths is suddenly about something?much greater numbers?of us know how to process and manipulate: words.?And this makes it so much more interesting to us, because our human language is such a large part of what defines us as human;?of what forms our humanity.

That’s why LLMs do such as good job of passing for human. An A.I. of this type is?not?a ‘digital human’, it’s just numbers doing cosplay.??

Mistakes/solutions?

But I don’t want to downplay the importance of A.I. by characterising it in this way, or to pretend that the feelings we have about it are not significant.

The process of?using tools like?ChatGPT,?Bard, etc.?which so many people have now done,?has led to a mix of powerful emotions: surprise, wonder, awe, excitement – but also a fair?bit of ‘uncanny valley’, and in the case of those extra fingers generated by text-to-image tools like DALL.E and?Midjourney, a strong?dose?of ‘the?ik’.?

A lot has been made of the?errors?and ‘hallucinations’ that generative A.I. tools seem prone to. It doesn’t seem all that clear, either, whether?these artefacts are just ‘teething problems’ which will soon be sorted out, or an inherent feature of the technology that we are going to have to learn to live with. This?could, incidentally,?be?a serious issue?with?the use of generative?A.I. tools for?learning.?After all, teachers,?trainers?and other learning professionals, if they’re to be of any use at all to learners, ought surely to be in the business of dispensing accurate information.?Where the A.I. is trained on a ring-fenced, authoritative set of materials there’s no problem, I am assured, but when it comes to generating?responses based on?uncurated?materials from?the open web, there could be limits to its usefulness.

But that’s not all there is to say about A.I. ‘mistakes’, by any means. Looked at another way, they?take on?a?really interesting?aspect. The?phenomenon of those extra?limbs and?physical distortions?you get when generating images of humans on?text-to-image tools?shows?A.I.making?mistakes a human would never make.?They show it’s clearly not sentient as some people have tried to claim, because it doesn’t know what it’s looking at. However, the fact that it can make such mistakes - a whole new type of mistakes, machine mistakes?– surely must mean that it can?potentially?come up with?very?useful?mistakes.

In the?human context we?often?call?useful mistakes ‘solutions’.?

A careless lab technician returns from holiday to find that mould has contaminated his petri dish –?and discovers penicillin. Mistakes, accidents, have been seen to be formidable generators of innovation down the ages. And A.I. makes the sort of mistakes a human would never make. The existence of a whole new category of mistakes, therefore, surely shows?acapacity to produce solutions, too,?that humans?would never come up with.?In that light,?ChatGPT’s?ability to summarize your company’s annual report in the style of William Wordsworth might not be the last word on generative A.I.’s creative potential.

Looked at in this light,?it’s?possible to see A.I. as a tool that can extend our creativity, by complementing our human creativity.?Replacing?human creativity, a big theme in the media, is probably much less likely.?The idea that in future A.I. will write all the symphonies, author all the novels, and throw the entire workforce of the creative industries in the scrapheap in the process, proceeds from a fundamental misunderstanding of what it is we look to those human artists to do for us. An A.I. trained on the entire corpus of human artistic activity can still only reflect the past. It can’t do the job of processing for us what is happening in this present moment we are passing through. Only another human can do that. It might give you music that sounds like Drake, but it can’t give you the next album by Drake, let alone the next Drake.

The true?creative?power of?A.I. might lie not in how similar it can make itself look to one of us –?but in its fundamental difference; its ability to make different mistakes, to find different solutions. But it will still need us, we humans, to recognise the value of those solutions.

It's a tool, not The Terminator.