Looking for a Found Curriculum: Regulating the Future – AI, Robots and Parliament

Several (many?!) years ago I was part of a course team that created a course titled “Robotics and the Meaning of Life” (we had also wanted the strapline “Joy, Fun, Robotics”, but that fell by the wayside when we didn’t secure a home experiment kit for every student).

One aim of the course was to introduce students to some of the technical issues associated with building, and controlling, robots: how robots move, for example, and the range of motions available to them; how they perceive the world; and how their perceptions can inform the decisions they might make in order to direct their motions to achieve some task.

The course also took a technological view in exploring a wide variety of social, legal and ethical concerns associated with integrating robots into our daily lives.

The course material lives on still, but is sorely in need of a refresh, not least in the examples we provide of contemporary robots. (That said, it’s surprising how well many still hold up as exemplars of sorts of robot you might find today.) Whereas twenty years ago, autonomous vehicles were the remit of science fiction, today they can be found on the public highway, albeit as test vehicles. And in the field of AI, whilst the toy examples we used in the original course still stand stand up in terms of demonstrating AI principles, there has been a shift from just demonstrating how emerging AI technologies might be used to living in a world where they are being used. Although we may not have noticed the shift. (It has been ever thus: one of the definitions of AI inside the field is that is stands for “Almost Invented”. As soon as the technique can be made to work reliably, it becomes just another programming library to be called on).

When it comes to updating the course, one way of understanding how developments in AI and robotics are evolving is to read the research literature. However, whilst it used to be the case that today’s academic research labs were where you’d go looking for tomorrow’s new products, many advances in AI and robotics are now driven by large tech companies and hopeful entrepreneurs.

The question of how to frame the updated course is also an issue: it currently exists outside of a qualification pathway in any meaningful sense, and does not set the scene for higher level courses on AI or robotics. It is essentially a standalone course. This means we are largely free to decide on the story we want to tell, and the topics we want to teach. But what would be useful?

For many technical subjects, there is often a canon of things to be taught and a recognised order in which to teach them. Several ostensibly different textbooks on the same subject tend not to be so different when you look at the contents list. This means as a teacher of a course, you often have a set of bases you need to cover. Introductory courses often have a bit more freedom associated with them but there is still a need to come up with some sort of guiding framework, or narrative.

The original Robotics and the Meaning of Life course was a short course released as part of John Naughton’s Relevant Knowledge program. The ethos of that programme always resonated with me: inform people in the way of the current technological world so they could put people right when discussing it in the pub. Which is to say: supporting a citizenry who are usefully informed about technology. Helping correct misunderstandings caused by our poorly developed “folk understanding of IT” (cf. folk physics or folk psychology) was another large part of it.

So might the update to Robotics and the Meaning of Life explore a new way of framing the material, or scaffolding the narrative? One of the problems with being able to access the world’s knowledge though a web search engine text entry box as an independent learner is knowing where to start. Taking a pre-existing course provides one way. My own self-directed learning tends to be motivated by trying to complete a particular task, which often leads to trying to understand any problem associated with it and the techniques used to achieve it. But what other models are there for “discovering” a relevant knowledge curriculum?

For the last few year’s, my general interest residential school evening talk has been titled “Who’s Driving the Future of Robotics?”. It’s based on numerous news stories and product announcements tracking the evolution of contemporary “real world” robots, interspersed with references to consultations, codes, regulations and (mooted) legislation that are required if the robots are to be allowed to become a part of, or enter, our society at scale. Part of the thesis of the talk is that young engineers should both track and engage with science and technology policy, because that will in part set the bounds on how technology will develop and impact on our society in the future.

So I wonder if the robotics course rewrite should be based around a found curriculum pulled together from UK Parliamentary activity around robotics and AI? The motivation, which I’ve half posted about several times before, is that after taking the course, a student should be able to make an informed decision for themselves about a technology policy matter and have a small amount of practical, technical understanding about that matter. They should also be in a position to debate the matter in a considered way and perhaps even contribute in an informed way to a related consultation.

The found curriculum itself would be bootstrapped from Parliamentary resources:

Government resources could also be called on. For example, consultation documents or outcomes such as on the Department for Transport consultation on Using advanced driver assistance systems and automated vehicle technologies. One of the advantages of making use of government consultations, and the responses to them, is that conflicting lobbying positions may also be identified as a way of exploring the topic in a socio-economic context.

Where technical topics are introduced, there would be an opportunity to do some technical teaching. This would be motivated by a desire to understand, or at least appreciate, at a technical and practical level what the technology involves and what its limits might be. For example, mentions of “deep learning” would set up a section on exploring in technical detail, albeit at an introductory level, what deep learning is and how it works.

Author: Tony Hirst

I'm a lecturer at The Open University, with an interest in #opendata policy and practice, as well as general web tinkering...