Significant work has been done in the past decade for digitization of ed curriculum. The Internet has been a major disruptor for educational content distribution with free, blended, and interactive models arising. The second wave came with gamification and big data analytics. Companies like Knewton began amassing enormous databanks of performance factors through online means to deliver adaptive learning.
This sets the stage for the next wave of educational innovation – the robo classroom. At Houghton Mifflin Harcourt, I prototyped adaptive learning integrated with augmented reality – much in response to Samsung’s demonstration of “the classroom of the future.” Samsung had demonstrated how a student could “paint” on a tablet, while a teacher could view the child’s screen and lock it whenever the child was supposed to pay attention. While the audience cheered at the “innovation,” I saw the dehumanization of education (See Ben’s post earlier). As a result, I wondered… what role can technology play that would advance a humanistic education. An I-Thou relationship as Martin Buber would say.
I built the prototype for the system. A teacher could scan the classroom to see each student’s performance (essential because adaptive learning means all students will be at different places all the time). This part of the development was easy. It’s just a dashboard with a camera. The next part chilled me.
I realized that as soon as an engine is built for adaptive learning an engine for adaptive teaching would appear almost as soon. This would chain teachers to a system, whether it be artificially intelligent or not. As a result, teachers become less empowered, innovation-starved, and further demoralized and underpaid/valued. WATCH THIS TREND. IT IS HIGHLY DANGEROUS.
I advanced the development to create an adaptive teaching engine. However, in doing so, I explicitly set the rule that the classroom teacher is the authority. The system is a guide, but the teacher is the expert. For example, an adaptive teaching system may know that Felisha learns best through peer-to-peer learning with Donna. However, it will not know that Donna’s parents are in the middle of a divorce and it may be more productive for an alternative path for education.
One benefit of an adaptive teaching system is that it provides CONTEXT for substitute teachers. And, if the data cloud expands to cover secondary factors (socioeconomic indicators, integrated government services like TANF or SNAP), social media, etc. The system can do a much better job of understanding the complexities of authentic learning.
When I learned IBM Watson was focusing on education, I was excited and worried. Excited because it is an interesting approach to education. Worried because I can pretty much guarantee they will do it wrong. In my discussions with IBM Watson Education leaders, it appears that they are simply replicating their work in Watson Healthcare, without considering that the models are completely different. I have played with the Watson system (read my blog for Watson Weighs in on the US Presidential Election) and it is crude. The ability for it to accurately extract semantics and identify emotions are at an infancy stage. It works best in a condition like healthcare and trivia (Jeopardy!) where the knowledge is highly structured and related. It does poorly in an environment where the data is unstructured, human-centric, and subject to interrelated fuzzy factors (e.g., socioeconomic, family, healthcare, DNA, weather, nutrition…)
I remember a presentation I gave where I closed by asking if technology will save us or destroy us. Using a google search, the answer was much higher that it will save us. But will we save ourselves from technology?