Sunday 26 June 2022

Learning, Dialogue and AI: Offline initiatives and Political Freedom

I'm running a small EU project in July called C-Camp. The idea is to instil and explore computational practices among students from 4 European Universities (Prague, Copenhagen, Milan and Heidelberg). I wanted to create something for it which built on my experiences in Russia with the Global Scientific Dialogue course (Improvisation Blog: Transforming Education with Science and Creativity (dailyimprovisation.blogspot.com) - about which a paper is shortly to appear in Postdigital Science and Education). 

In Russia, the vision was to present students with a technological "cabinet of curiosities" - a way of engaging them in asking "this is interesting - what do you make of it?". It was the uncertainty of encounter with technological things which was important - that was the driver for dialogue, which dominated the course. C-Camp is very much in the same spirit. 

This time, I have been a bit more ambitious in making my cabinet of curiosities. I've made a cross-platform desktop app using ElectronJS which incorporates a tabbed web-browser, alongside self-contained tools which make available learner's data to the learners (and only to the learners). The advantage of a desktop tool is that, apart from the learners being able to change it (my programming and design is merely functional!), nothing personal goes online, apart from the traffic in each website.  The data of engagement with the tools - which is something that is usually hidden from learners - then becomes inspectable by them. There  are lots of "cool tools" that we suggest exploring (like the amazing EbSynth below)

The pedagogy of the course will then be to explore the data that learners themselves create as they process their own uncertainty. It's messy data - which can be an advantage educationally - but it illustrates a number of important principles about what is going on online, and what data big tech companies are harvesting, and how they are doing it. 

More to the point, by having a desktop tool, there is an important thing to say that "edTech doesn't have to be like the LMS!". Not everything needs to be online. Not everything needs to be harvested by corporations. And more to the point, if individuals were more in contact with their own data - particularly their own learning data - there are opportunities for deepening both our learning and our engagement with technology. So supporting students in downloading and analysing their own Facebook data can be part of a journey into demystifying technology and inspiring the imagination to look "beyond the screen"

 


One of the things I've done is to integrate 2 AI services. One of them uses the OpenAI service, which is online. The code for doing this is quite simple, but the important thing is that the processing happens remotely on OpenAI's servers. 

However, the other AI service is local. I've integrated the VGG16 model with Imagenet data so that students can upload and explore image recognition. The model and the code are all on the local machine. The point to make is that there is no reason why OpenAI shouldn't work like this too - other than commercial reasons.

What fascinates me about this is that for all the anxious talk about AI and its supposed "sentience", nobody talks about the technical architecture which basically up-ends the idea that everything has to be online. Large-scale language models are basically self-contained anticipatory dialogical engines which could function in isolated circumstances.

Think about this: imagine in a non-free country like Russia or China, where the authorities seek to monitor and control the conversations that individuals have, suddenly individuals can have conversations which are not monitored - simply by being in possession of a particular AI file. 

I'm doing a demo of OpenAI tomorrow in China. The last time I did it there, it worked. I doubt it will work for much longer. But it's easy to envisage a future where a market for specialised language model AIs start to infiltrate the underworld allowing people to have "prohibited conversations". That could mean both very good things for social organisation and freedom from oppression, and bad things in terms in terms of crime. 

That is one of the more fascinating things to discuss in C-Camp. I think I might be more careful with my Chinese audience!




Saturday 25 June 2022

How Learning Feels

When learning works, it feels like a burst of energy. It is the energy of an explosion of new possibilities brought about through some revelation. It is a spiritual moment (something we hardly ever acknowledge) - even when it is learning about unspiritual things. Like the discovery of a new physical energy source, we can live off the energy of new learning for some time. 

Striving for this moment is not easy. Yet we are driven towards it for reasons we do not understand. Teachers often assume that the motivation is produced by the mere operation of the education system. But the education system exists because curiosity and the motivation to learn exists. The system has no explanation for curiosity, and it struggles to conceive of ways of learning outside of itself.

New possibilities are possibilities for new social action. It is not just what some sociologists call "agency", but a transformed social configuration. A learnt skill is a transformation in social connections and conversations. It is new dialogical potential. And dialogical potential begets new possibilities for learning and energy distribution among others. To talk of the energy of learning, we should also talk of the energy of teaching. There is an energy flow in these dynamics.

In natural ecosystems like ponds and meadows, energy dynamics are very important. Ecosystems maintain themselves by keeping the energy flowing between co-evolved co-habiting system components. If the flow is stopped - by environmental damage, for example - the ecosystem dies. 

Education systems have become tragically good at preventing flows of energy. Instead of allowing energy to flow, education systems hoard it, exploit it, seek individual gain from it, use it to make money, and seek to make ourselves "powerful" as if we are independent from everyone else. 

We do this partly because we do not understand the dynamics of energy. If we did, we would take music much more seriously because it is one of the few human activities which exhibits energy flow in a pure form in a human system.

Intuitively, I think we know this. It is a symptom of the education system that it prevents us from "knowing" what we know deep down. Somehow we need the education system to adapt so that  it helps us to steer ourselves through what we know deep down. It needs to ease our steering - particularly in uncertain times. It is a transformation from hoarding knowledge to assisting steering. Then perhaps the steering of learning will feel more natural.


Wednesday 8 June 2022

Trimtabs and Loosening Creativity

Creative processes are often difficult. It is hard to steer through distractions, uncertainty, self-doubt, dead-ends, etc. The steering becomes "heavy". So what's wrong with the mechanism, and what might be done to loosen things up to make the process more navigable?

The construction of niches for creative work is critical. It is the niche within which new things can grow. From a technical/theoretical perspective, niches are the result of redundancy. In his description of the Zone of Proximal development, Vygotsky said as much (without using the word "redundancy"), in highlighting the importance of imitation in what he called the "learning" process, and arguing that "development" lags behind "learning". In the same way, creation lags behind redundancy - it doesn't matter what kind of creation it is - it can be technical, artistic, organisational, theoretical or scientific. 

Margaret Boden talked once of the creative work of Spanish seamstresses making Flamenco dresses. She said "they do one layer, then another, then another, then another... what's going on there?" It's the same with things like mosaic, quilt-making or knitting. I didn't know enough about redundancy at the time to suggest it as an explanation, but I think she was already thinking this. This is niche construction. 

It is something we tend to ignore in education because we have become so obsessed with outcomes and products, seeing the processes which produce them as "problem solving". The word  "solving" is interesting because it really means "loosening" - solvere. That's not how people who talk about problem solving think about it. But if loosening really happens, then it makes the "steering" easier.  

Buckminster Fuller's idea of a trimtab is a loosening device. It literally loosens the steering, and it does it by creating a niche for steering - simply by  adjusting the pressure on a rudder or a wing. This tiny thing at the back-end of the navigation process is the thing that makes everything else work. Now perhaps its not stretching things too far to say that trimtabs create redundancy. Without them, there are a variety of different forces and pressures operating on the wing - so much variety that there is no single steering movement that can manage this variety. The trimtab reduces the variety by increasing the constraint. It's rather like a spider spinning a web. By creating a uniform area of lower pressure,  steering can be assisted. 

The trimtabs of our organisations lie in the redundancy of communication among their workers. Where there is high redundancy, we will also see what we might call "collegiality". Collegiality, team working, and a shared mission can all create the niche for organisational creativity. An absence of it will make creativity very difficult. 

Our organisations do not have operational trimtabs. The only lever they can pull is the departmental meeting - and this has become a ritual which often serves very little purpose. There is a deep need for exploring new mechanisms for institutional organisation. The answer to this lies in technology - but not the kind of surveillance technology which is often talked about (like "learning analytics"). Surveillance will not produce collegiality. Quite the opposite. 

We need to use technology to provoke dialogue among colleagues. It is through the dialogical engagement among colleagues that effective niches can be established. This is not to see technology as instrumental, but dialogical. AI may be our best opportunity to do something like this, and if there is one single challenge that faces us with that technology, it is that we misuse it to tighten, and not loosen, the steering.   

Saturday 4 June 2022

The Cybernetics of the Trimtab Society

Over the last seven years, I've been heavily involved in a medical diagnostic project which unites human and machine judgement. This has always been cybernetic in my mind (and it was cybernetic insights which led to some pretty cool machine learning that sits behind it). It's about to be commercialised which is very exciting, not least because the technology is applicable to fields far beyond medical diagnostics - education, management, organisational risk and public health are all within scope of potential application. 

Cybernetics relies on simple rules and metaphors, but these work in a wide range of contexts. The Law of Requisite Variety is the most important - the amount of variety (or complexity) that a controller has is the limit of the complexity of any system that it can control. Most simply, variety eats variety. Since most systems have to survive in environments of greater complexity than they possess, they must establish a controlled relationship with their environment through attenuation (selecting what information to pay attention to and what to ignore) and amplification (use their capabilities and understanding to create a niche in the environment - for example, a spider spinning a web). This can balance the variety equation.

A simple mechanical metaphor of cybernetics is the Watt Governor on a steam engine. The engine's speed, represented by the spinning of its flywheel, is controlled by a device (the governor) which uses centripetal force generated by the speed of the wheel to either slow down or speed up the flow of steam to the engine. This works because the wheel has exactly the same amount of variety as the governor: whatever state the engine is in is matched by a corresponding state of the governor.

This is fine as a metaphor, but in social life, there is no one-to-one mapping of environmental complexity to controllers, so we end up with very complex patterns of attenuation and amplification which can create dangerous positive feedback to the system. We are living through this in many ways at the moment - not just in the climate crisis, but in the political feedback from our online communication, the economic system producing runaway inequality, the Ukraine war, and so on.

Buckminster Fuller drew attention to a different kind of cybernetic feedback mechanism - the trimtab. Trimtabs are the small edges on the back of wings which wiggle as the plane is flying, and which serve to make the pilot's job of steering and stabilising the plane easier. In other words, the trimtab is part of a mechanism which connects the pilot to the machine. It is not self-enclosed like the Watt Governor, but translates the environmental conditions into a potentially controllable situation, which would otherwise be very difficult to control. 

Buckminster Fuller thought so much of trimtabs that he had "Call me trimtab" written on his grave. He argued that the most important part of steering was not at the front, but at the back, and that each of us could be part of a "social trimtab" each feeding information about environmental conditions in a way which could facilitate effective steering. 

The diagnostic AI which I and our team have created basically works like this. With our work, the "pilot" is the doctor, but the pilot's job is to steer through different environmental conditions in terms of differing degrees of prevalence, diagnostic certainty, organisational complexity, health economics, risk and potential positive feedback. To achieve this has entailed a very different approach to AI. Conventional AI is simply used to provide "answers", often with the intention of replacing the "pilot". That's not a good idea because it throws away huge amounts of information which can be critical to understanding the nature of the challenges we face. The trimtab (and our trimtab AI) by contrast preserves information, transforming complex data into the conditions wherein effective decisions can be made. 

I've always felt that the most important thing education should do is to harness the uncertainty of individuals, because this information is information about the nature of our environment. What I've never been entirely clear about is how this "harnessing" looks - lots of forums, debate, etc, don't seem to work and in fact amplify social complexity. So we need a way of organising the many different signals coming from society as a means of facilitating effective steering for the planet (or Spaceship Earth as Fuller said). This may be the most powerful and effective use of AI.