Teachers vs Robots

A panel presentation at the 2017 meeting of the Information Architecture Summit

On March 24, 2017, I presented the talk Teachers vs Robots: Information Architecture and Assistive Artificial Intelligence to the Information Architecture Summit in Vancouver, BC, Canada. The talk was based on one component of my work for Eli Review in which we designed information structures specifically so that our software could watch for and help teachers make evidence-based decisions in the classroom.

The original slide deck is available on Google Drive. Below are sample slides, full video from the talk, and the original proposal.

Photo of me presenting at #ias17
One of the slides from my slide deck - here, I describe the purpose of the machines we build, which is to assist rather than replace human teachers.
Another slide - here I discuss how we arrived at our learning indicators by starting broad with theory, winnowing down to indicators, then building detectors.

Session Title

Teachers vs Robots: Information Architecture and Assistive Artificial Intelligence

Short Description (300 words)

Information workers face being replaced by artificial intelligence just as many types of workers face being replaced by robots. We are told that machines can work faster and better and cheaper than humans, and in some ways they already are: robots are writing simple articles for the Associated Press, for example. There is still time to make the case for humans, however, and information architects are uniquely positioned to lead in the development of ethical technologies that augment and empower us rather than replace us.

This session will illustrate the role information architects can play in advocating for humans by presenting a case study in the development of a learning technology. Writing teachers face a serious threat from machines capable of scanning student essays and giving automated feedback – machines that can seemingly replace them. The speaker will make a case for why these machines can’t effectively replace human teachers and demonstrate how he helped build an app with teachers that actually had the potential to help students improve. He’ll illustrate how information architecture guided that process from requirements gathering to deployment and how understanding teacher data needs made it easier to develop tools to augment their work. With machine assistance and well-designed displays, the app could provide teachers with analytics and insights about where they could make timely, precise, evidence-based interventions that would have the greatest impact for students.

From this case study, the speaker will offer some generalized principles for information architects that can help them study human knowledge work and guide the development of ethical machines to augment that work. Information architects are best qualified for understanding and describing the relationships between humans and information, and if we don’t work toward supporting and augmenting humans, it won’t be long before we’re working for – not with – our robots.

Twitter Description (140 characters plus @mcleodm3 and #IAS17)

Teachers vs Robots – @mcleodm3 discusses how useful, actionable information about learning can make human teachers indispensable #IAS17

Long Description

This session will mostly be stories I have from developing a learning technology over the last five years. Specifically, I believe they’re of interest because the tech came from a classroom – we needed data for our instruction, but we tested an analog model that depended on three PhDs and two master’s students manually gathering data and updating a spreadsheet. From there we built a prototype and then a full commercial app, all based on this manual data gathering exercise.

I’ll share a lot of specific info about our object models – what objects are valuable, how we knew they’d be valuable, what they look like, and how they change over time. I’ll share stories of what we learned from teachers during usability tests that shaped how we gathered and displayed that data. I think this will be interesting because I can show both database structures and interfaces and discuss how they two evolved together based on user needs.

I’ll also share specific stories of failures and where our assumptions about what data would be useful turned out to be completely wrong and where we pivoted to deliver something that actually was useful. I’ll also discuss surprises we had about how teachers use both the quantitative and qualitative data our app provides that led us down paths we never expect. This should be interesting because teachers aren’t often presented with learning analytics like ours and we can provide specific examples of how teaching has changed as the result of having data early in the course when they can act on it.

Lessons / Takeaways

  • A specific use case for how teachers use information technologies
  • A model for participatory design with teachers and students
  • Detailed examples of an evolving data model (what data to gather)
  • Detailed examples of evolving data displays (what the user needs and when)
  • An example of ethical design – machines helping rather than replacing humans

Why are you qualified to speak about this?

My background is in education and technology, specifically writing education. I have two graduate degrees, one in English language and literature and one in digital instruction.  I have eight years experience teaching writing, technology, web design, and user experience at the university level. I worked for five years at a university research center from which my colleagues and I spun out an edtech startup. Our startup’s first product, Eli Review, is the one described in my proposal; I’m responsible for the product’s design and development.