Until recently I was a high school science teacher in Texas. I taught at an alternative high school [1] that focused on dropout recovery. I think teaching there gave me a glimpse of the future. Because the kids coming to us all had bad experiences of school, had failed various classes and had often already given up, we taught each of them individually.

They learned at their pace — sometimes slowly and sometimes faster. Some finished all of high school chemistry in three weeks, some took more than a year. I modified the curriculum as they went, drawing on different assignments to match their strengths and weaknesses and tailoring each to their needs and interests. At the same time, I mentored them individually on social skills, life skills, further education and careers. It worked, we helped more than 200 kids graduate each year who otherwise would not have done so.

We could do this because we had tiny classes — up to 5 or 10 kids. Do the math: a 60 minute class with 10 kids gave me 6 minutes with each child. It’s not much, any more kids and it became a horrible mess. And, the teaching budget per child in our school was three times that of the regular high school.

You may have heard about the promise of intelligent tutoring systems that can individualize feedback and instruction for each student and could give all kids a similar experience as at my school, even in larger classes. We need these tools — our schools fail far too many of our kids today.

But there’s a problem and it’s not the technology.

We can already infer confusion, anxiety, boredom, pleasure, excitement and a host of emotions through video input, mouse clicks and other sensors including smart watches and brain wave scanners [2]. We can already model kids’ cognitive processes, domain understanding and interests in data sets [3]. We can already dynamically generate educational material in text, graphics and animations and provide feedback through personalized agents [4] and haptic feedback [5]. We can also share these profiles across different subjects and over many years [6]. We just need to put the pieces together.

“Gartner named [adaptive education] as the top strategic technology for higher education IT leaders to plan for in 2016 because of its potential to help scale personalized learning” [7]

But there’s that problem.

That problem is the collection of vast amounts of very intimate data about how our children think without first securing how it should be stored, collated, monetized or shared. We need to solve this problem first and it’s one of those problems that we won’t care about until it’s too late.

If these teaching systems are to work properly they need to understand how each child processes new information. How they fit it into their existing knowledge. How they cope when it doesn’t fit into their mental models. Whether they are strong in math, logic, reasoning, empathy, imagination, critical thinking etc. These models will be used to focus instruction and tailor lessons to kids’ needs so that they get a better educational experience. But we’ve seen what Cambridge Analytica could do with a simple personality survey [8]. Imagine what could be done with 20 years of continuous observation of kids while they learn to become critical thinkers?

Imagine what could be done with 20 years of continuous observation of kids while they learn to become critical thinkers?

The solution to this will be a combination of technology, regulation and enforcement but it won’t just emerge — we, you have to drive it. Currently in the US we have COPPA to protect young children’s data [9]. It prohibits the commercial use of data from children under 13 without parental consent. But without enforcement, it does not work. In April this year a study found that out of the nearly 6000 apps on the Google Play store aimed at children, 57% appeared to be violating COPPA and each had been installed at least 750,000 times [10]. And as far as I know, they’re still there.

So — we have a problem that needs solving before it becomes a problem. The benefits that technology can bring to education are so exciting but someone is going to know how just what makes us tick. How we think. We need to figure out how that can be made safe.

References

[2] For a review of technologies (both through sensors and through analysis of user interactions with the software) for detecting affect and emotions in students using online instruction See DeFalco, J. A., Rowe, J. P., Paquette, L., Georgoulas-Sherry, V., Brawner, K., Mott, B. W., Lester, J. C. (2017). Detecting and Addressing Frustration in a Serious Game for Military Training. International Journal of Artificial Intelligence in Education, 1–42. https://doi.org/10.1007/s40593-017-0152-1

[3] Santos, O. C., Kravcik, M., & Boticario, J. G. (2016). Preface to Special Issue on User Modelling to Support Personalization in Enhanced Educational Settings. International Journal of Artificial Intelligence in Education, 26(3), 809–820. https://doi.org/10.1007/s40593-016-0114-z

[4] Johnson, W. L., & Lester, J. C. (2016). Face-to-Face Interaction with Pedagogical Agents, Twenty Years Later. International Journal of Artificial Intelligence in Education, 26(1), 25–36. https://doi.org/10.1007/s40593-015-0065-9 ; Kim, Y., & Baylor, A. L. (2016). Research-Based Design of Pedagogical Agent Roles: a Review, Progress, and Recommendations. International Journal of Artificial Intelligence in Education, 26(1), 160–169. https://doi.org/10.1007/s40593-015-0055-y

[5] Escobar-Castillejos, D., Noguez, J., Neri, L., Magana, A., & Benes, B. (2016). A Review of Simulators with Haptic Devices for Medical Training. Journal of Medical Systems, 40(4), 104. https://doi.org/10.1007/s10916-016-0459-8

[6] IEEE Learning Technology Standards Committee (LTSC) — Assuring Compatibility Across Ed Tech Product Categories — http://sites.ieee.org/sagroups-ltsc/home/

[7] Adams Becker, S., Cummins, M., Davis, A., Freeman, A., Hall Giesinger, C., and Ananthanarayanan, V. (2017). NMC Horizon Report: 2017 Higher Education Edition. Austin, Texas: The New Media Consortium. Available at https://www.learntechlib.org/p/174879/report_174879.pdf

[10] IMDEA NETWORKS INSTITUTE. (2018, April 24). Thousands of mobile apps for children might be violating their privacy. EurekAlert! Retrieved from https://www.eurekalert.org/pub_releases/2018-04/ini-tom042418.php