Knewton, once hailed as a “mind-reading robo tutor in the sky,” is no more, having been sold for parts to publisher John Wiley & Sons.
As reported in the Chronicle, Wiley plans to use whatever scraps are left over as part of its low-cost digital courseware offerings. I don’t know what that means – perhaps there is software that is useful, and I bet there’s lots of student data - but it’s quite the comedown from what Knewton founder Jose Ferriera promised in a 2015 story at NPR.
"We think of it like a robot tutor in the sky that can semi-read your mind and figure out what your strengths and weaknesses are, down to the percentile."
Even at the time, there was much skepticism over Ferriera’s claims, including that from longtime ed tech observer and consultant Michael Feldstein who said that Ferriera was selling, “snake oil.”
I will admit that as a critic of adaptive learning software as it is being used in education, and a particularly big critic of the hype surrounding these things, I experience a certain schadenfreude at this news. My first caution about Knewton in this space dates back to 2013 when I flaggedFerriera’s stated goal “to create individual, psychometric profiles that would presume to say, with statistical authority, what students know and how they learn” as both wildly fantastical and hugely intrusive on student privacy.
Often, even if the technology worked exactly as advertised it looks corrosive, as with the latest trend to track “biometric” inputs to see if students are “paying attention.” Attention is not a proxy for learning, but because it’s what can be tracked, we treat it that way.
This is not good because learning is more complicated than mere attention. The technology can only be effective if success is defined down to the thing the software can do. Apparently, in the case of Knewton, that’s not nearly enough.
This is a smaller story than the recent “rebellion” in a Kansas school district over Summit Learning’s adaptive software platform, but we should notice some commonalities in that both Summit and Knewton (in its heyday) were “free” for schools to use. It is not accidental that students at under-resourced schools are subject to these experiments.
The software is always packaged as an aid to teachers to “free them up” to do the important stuff, but I also hope people are growing wiser than to accept these claims at face value.
In my field of writing, for algorithmic grading software, it is simply a lie, as there is no more important thing a teacher of writing can do than read and respond to their students’ work. As I say in Why They Can’t Write: Killing the Five-Paragraph Essay and Other Necessities asking a writing teacher to work with students when they know the grade but haven’t read the writing is like asking a football coach to work with their team when the know the score, but haven’t watched the game, and orchestra conductor who has heard the applause, but not the performance itself.
It is nonsensical, and yet many dollars are being invested in creating algorithms that “grade” student writing. The only possible outcome is to load teachers with more students, a move that is the opposite of what we should be doing if we want to improve writing instruction.
Even in other fields, the goal of the adaptive software is to replace the teacher. It’s prevalence in under-resourced schools shows that this is already happening and reinforces the notion that we need not re-invest in these schools and communities because: software. It is corrosive. It should be resisted.
Yes, #notalledtech. I know that there are conscientious engineers working on creating genuinely meaningful education technology applications, and I hear from them every time I write a post like this. Consider yourselves exempt from the critique, but also, you should be leading the charge against those who are promising mind-reading robo tutors and those who think a mind-reading robo tutor actually sounds like a good thing to begin with.
The truth is, I can’t take much pleasure because I think about the time students have wasted being experimented upon. Arizona State initiated a “full deployment of Knewton, for both its online and blended courses in 2011.” Pearson was full-steam ahead with Knewton, putting it in front of students, until it stopped. What did those students experience? What are the consequences?
More than $180 million in venture capital funding was sunk into Knewton.
For the ed tech edupreneurs, failure only means a chance to move on to the next project.
Ferriera is heading up something called Bakpax, which judging from the website, is platform where you can take a picture of your math homework and have it instantly graded, or something.
On the other hand, students only get one shot at each class, each grade. Some of the students in Kansas subjected to Summit Learning’s software got hand cramps, and headaches for their trouble.
The saddest thing is that I could write some version of this post every week, and yet nothing in the big picture is going to change.