Ai is Everything That is Not Needed in Education

I guess I will start off with the blogger cliche: it has been a while. I know. It’s not that I don’t have anything so say – it’s just that it feels repetitive to keep talking about Ai. It never really improves – not in any true way. And the news just keeps getting worse and worse.

Every day I could write posts pointing to the daily articles about something going wrong with Ai. “Microsoft Scales Back Ai Goals Because Almost Nobody is Using Copilot” says ExtremeTech. Told you so. “Investors Expect Ai Use to Soar. That’s Not Happening” says The Economist. Again – not a surprise.

(Some of the stats in The Economist article are telling. When asked in a survey “Do you use Ai on the job?”, 87% of the executives surveyed said yes, 57% of the managers said yes, and only 27% of the workers said yes. The people actually doing the work aren’t finding Ai useful. Also not surprising from what I am hearing from workers in various sectors. As many have said in different ways: Ai is not coming for your job. CEOs / employers / upper level management are using Ai as an excuse to replace you with a sub-quality replacement.)

To be honest, it gets too easy to pick on Ai. It was overhyped by people who knew better.

But since it is still being forced into education in so many ways, I still have to pay attention. I wanted to go back and see what is happened with Matter and Space, an educational Ai company that claims to be “redefining what it means to learn in the age of Ai.” I have expressed concerns about Matter and Space in the past (as have others). I looked at their website a few months ago and saw several changes to the website, seeming to indicate they had gone live with their “Learning Environment 1” (LE1) software. I formulated some responses to what I saw there and then life got in the way. I went back today and found that the Matter and Space website redirects to the Southern New Hampshire Universities’ main website. Huh.

There is still an archived version of the Matter and Space website on Archive.org. This is final version of it, with the released version of LE1. It is still unclear what LE1 is exactly. The problematic video Butterflies appears to have been removed from the front page at least (I still found it at the bottom of a subpage, but the Way Back Machine doesn’t appear to have saved most of the subpages). There is still a video on the front page that contains many of the statements that concerned me the most (you can still view it here).

Paul Leblanc still asks what learning could look like “if we were unconstrained by the way it happens today?” But then the website is filled with all kinds of terms that are constraining it today, like skills, competencies, learner success, engagement, mastery, persistence, real world readiness, etc.

George Siemens still says that they want their Ai to know learners better than any other educational system has known you in the past (Creepy. Probably something white guys don’t worry about as much as everyone else). He also says “the best established pedagogical model for how students learn is a question and answer type of dialogue.” The word “best” is doing a lot of lifting there – lifting that is not backed up by research. But then again, it isn’t clear what is meant by “question and answer” here. Does the learner ask the questions? They usually want an education because they don’t know what questions to ask. Does the instructor/Ai ask the questions? Isn’t that just… testing?

The video still shows what is basically a chatbot app. Siemens then lists a lot of stuff that LE1 is going to give learners (content, resources, material, social connections, a nosey Ai bot that will pry into their lives to get to know them better than any before, etc) that goes well beyond “question and answer” models. Does that mean LE1 uses a lot of stuff that is not “the best established pedagogical model”?

Tanya Gamby says “We can see where you’re struggling in real time. We can slow things down, let you take a break. Education’s never existed like that.” Except that, yes, every in-person class K-16 allows teachers to see that students are struggling and let them takes breaks. That has happened to me thousands of times in my life. I really hope that was just a bad video edit and that she doesn’t mean to imply that millions of teachers have never given wellness breaks.

Now that Matter and Space website seems to be gone, I wonder what happened. All I can really find online is a LinkedIn post by Paul Leblanc, who says that Southern New Hampshire University has decided to bring the platform in-house. LeBlanc does not appear to be staying with the project, and announces he is moving on to the Harvard University Graduate School of Education as a visiting scholar and special advisor. Oh, and he is writing a book on Ai and education.

I can’t find much else on what has happened. Back in September, Siemens wrote a blog post called “What I Learned Building an Ai University Over the Last 2 ½ Years: Part 1 of Many.” But there have been no more parts. It seems like things were looking good while this was written. But even here there is much I disagree with when it comes to Ai in education.

The post starts off with the usual concerns for Universities not responding to Ai. That has not been my experience, considering we already have policies and degrees and tool integration that are already a year old or more in universities every where. But regardless of any of that – no one has ever stopped to prove why universities have to respond at all other than “because it exists!” Even Ai Cheerleaders will tell you they don’t know where Ai is going exactly, what exactly all is going on behind the hood, or even how Ai companies can protect their users from harm. Universities have these pesky rules about harming students – so why should any University adopt any tool where the main response by the companies’ owners to user safety issues has been a mostly collective shrug?

Students, faculty, and employees of universities are people, and we really should care enough about people to leave behind the “move fast and break things” mentality.

Siemens says things like “higher education faculty and staff need to become Ai product builders” and “learning as an act itself will be massively augmented and improved by Ai” that all sounds like past trend hype as well (“educators need to be bloggers / build MOOCs / create islands in Second Life / understand crypto / master learning analytics / etc / etc / etc” and then “learning will be massively improved by blogs / MOOCs / virtual worlds / crypto / analytics / etc / etc / etc.”

People are finding Ai-chatbots annoying. Research is starting to find that Ai is harming critical thinking skills rather than helping it. Even if someone could find good news about it, all of the environmental and social harm it causes would still outweigh any benefits.

Siemens refers to LeBlanc as “aggressively impatient,” driven, and decisive. He seems to think this is what education needs more of. I would disagree. Education seems to have had plenty of that for decades. Education is often a slow process that takes time to build trust and community. You have to go back and repeat things all the time. You have to take the time to be mindful of the process. We know what works in education: time, funding, nutrition, safety for all, etc. We know that our politicians and leaders are unwilling to fund what works – therefore they so desperately need Ai to fill gaps it is never going to be able to fill. That way, they can keep money in the hands of their big business cronies, not flowing back to the people that paid for it in taxes in the first place. Ai (poorly defined as it is, but in it’s current form) is everything that is not needed in education.

Are You Ready for a Trip into (Ai) Matter and Space?

Another week, another new round of Ai start-ups. This week we actually get one that is focused on education called Matter and Space. Full disclosure: I read about this one first through Audrey Watters post Matter and Dystopia, so go ahead and get those accusations of bias and not understanding what I am talking about ready if you must. I will probably repeat some of her points, as well as others that have responded.

A quick tour through the website reveals a slick (aka expensive) design along with even more slick (aka even more expensive) videos. We see a lot of claims of doing things different… and then a three person leadership team that is all white and 2/3 male. Not a good way to start off a new way of doing things…

The first thing you see on the website is a short film called “Butterflies” featuring Lakota children on the Pine Ridge Indian Reservation, as well as the first boy on that reservation to get a slicker new Google Glass-looking device. If you are like me, your first thought is “I wonder if they thought about how it looks to have a white-led team out to ‘fix’ indigenous education?” Well, if you stop the video briefly at the 6:07 mark, you get a long statement about the harms of boarding schools, and how the project seeks to be better. It is only up for 3 seconds, so it is easy to miss. Especially because it shares a 3 second slide with a land acknowledgment (even though indigenous leaders have spoken out about making those only performative) and an acknowledgement that Ai is destroying resources. But that is someone else’s problem – they call on others to fix it.

It would seem that if they wanted to do something different and make sure their acknowledgments are not just performative, then they could have at least put native educators in charge of the company? Or even hired an indigenous person to write and direct the short film? Or made the acknowledgments more prominent and easier to read? I wonder how much the people in the film knew about what they were participating in, since all of the Ai interface and voice acting was obviously inserted post-production?

Why Matter and Space?

So who are these three leaders of Matter and Space? For full disclosure, I am not familiar with Tanya Gamby, so I don’t have a  fuller context to place her words in the intro video into. I used to work for George Siemens at the LINK Research Lab. It seems like the motto of Matter and Space (“Human-centered learning for the age of Ai”) is the next evolution of LINK Lab’s motto (“What does it mean to be human in a digital age?”). There is an interesting shift there from asking an open question that anyone could answer differently to stating that they will be the ones providing the answers.

(Random ironic memory: Siemens was famously ridiculed for wearing Google Glasses at a conference circa 2013, creeping out the attendees who had privacy and safety concerns. Matter and Space starts with video that shows a Google Glass clone with even creepier abilities than Google Glass being used on children. Indigenous children at that.)

I have heard of Paul LeBlanc and the controversies that involve him. He spoke at a conference I was at a good 15+ years ago. I can’t remember if he was the keynote or on of the major sponsors that these conferences often let speak for almost as long as the keynote right before they start. I just remember he said a lot of things that people liked, but if you thought about them for a second, they started sounding concerning. He continues that pattern in the “Why Matter and Space” video.

For example, early in that video, LeBlanc asks “what could learning look like in the age of Ai if we were unconstrained by the way it happens today?” Good luck becoming unconstrained from politics, budget shortfalls, conflicting parental rights movements, and all the other things education currently wishes it could also become unconstrained from as well. Plus, beyond all of the constraining stuff, there is also a complicated layer of what “happens today” that is there to accommodate different community cultures, different learner needs, different parental goals (including religious ones, or lack of religious ones), and so on. There are “constraints” designed to protect individuals that are also very important constraints. Even though they are currently being eroded or tossed aside in some countries. Trump also often talks about getting rid of education how it happens today. Again I ask: shouldn’t it cause people some pause when they end up sounding like a politician like him?

Siemens is up next saying “our goal with Matter and Space is really to know our learners better than they have ever been known by an educational or really any other institution.” Well, institutions can’t really “know” anyone – the people in those institutions can. But anyways, an Ai also can’t “know” anyone, even though he later states “an Ai agent is going to get to know them well.” Generative Ai can predict a next likely response to an input from an individual, but it doesn’t know that person on any level, or even if it is correct or not. When it is correct, that came from prediction, not “knowing” someone. Plus, if they “hold deep respect for the privacy and personal autonomy of the individual and their data” – how do they think their Ai will get to know a learner better than anyone else has? You have to give up a LOT of privacy and data to be known by another person. You can’t really know someone well if they want to keep information about themselves private.

(Side note: do you remember when people were trying to come up with a “better” term for Learning Management Systems? One of the jokes for a bad idea was always “Learning Operating System” – which seems to be used seriously here?)

Next we have Siemens stating that “the best pedagogical model for how student learn is a question-and-answer type of dialogue.” I’m not sure what he is referring to here. It is a stretch to call this kind of dialogue a full, complete “pedagogy,” but yes it does work well in some contexts. But not others. Do you really want someone to learn heart surgery, engineering, cooking, or any number of jobs primarily by question and answering? One of the first things they teach you in a Ph.D. program is to get rid of absolute statements that utilize words like best, worst, never, always, etc – because research is rarely that black and white. We actually don’t know as much about how people learn as many would like to imply, so stating which one is “best” is difficult as best, fruitless at worst (because it would depend on context, intention, etc).

Siemens goes on to describe a system of “content, resources, materials, social connections” that is not really question-and answer dialogue most of the time… and sounds just like parts of course design that would cost a lot more than zero. Is this the answer to how course design would cost zero? Because this is a very expensive proposition.

Gamby states that “we can do what we call ‘just-in-time support.'” (Does she think they came up with this idea that has been around for decades? That second ‘we’ in there is a bit odd…) “We can see where you are struggling in real time. We can slow things down, let you take a break. Education’s never existed like that.” Never?!? I did that yesterday. I did that all the time back when I taught 8th grade: stop and talk with students about what they are struggling with, and then recommend they come in for tutoring time specifically designed to do what Gamby claims has never happened. I do that now in online courses all the time – giving students more time to complete assignments while talking with them about what they need a better understanding of. I would estimate that this education she claims NEVER happened actually happens millions of times daily in the United States alone. Does it always happening when it needs to, or for all students that need it? Nope. But to say that “education’s never existed like that” is incredibly inaccurate.

Siemens says “we focus on the entirety of a person, not just their knowledge and skills”… but does their Ai really do that – or someone else? LeBlanc talks of tired teachers in classes of a hundred kids, and asks with a straight face “what if we could equip every child in that crowded classroom with a device?” He speaks as if he is the first to discover ‘one device (laptop, Chromebook, phone) per student’ initiatives. But then he talks about this technology freeing up teachers because they know the students better. This in turn frees up the Ai for “knowledge transfer.” Wait – is it the Ai or the teacher that knows the student better? Does the Ai focus on the entire person, or is it the teacher that does that? Knowledge transfer isn’t knowing students. Knowledge transfer isn’t even a good way to understand pedagogy. It’s a horrible term to use if you claim to want to do education different or even well.

(When people mention “knowledge transfer,” I always want to stop and ask if they even know that the “brain is a computer” metaphor hasn’t been considered invalid for a long time. Or that the way people “transferred knowledge” in The Matrix was pure science fiction. But usually I just say “I know kung fu” to myself and write-off who ever used the term.)

When you look at state level education standards, at college level business goals, and bigger pictures of institutions, you always find statements about how these institutions currently “focus on the entirety of a person, not just their knowledge and skill.” Look at all the whole person-type classes they teach to create “well-rounded students,” at how many clubs and extra curricular activities and mindfulness classes and so on and so forth that they offer. Not always well, but still… It’s not really new to focus on the whole person. Traditional education has done that for a while, even though often falling short.

Honestly, all of this sounds a lot like “education the way it has been done before,” but with a wave of the magic Ai wands to somehow make it look different.

What is Matter and Space Learning?

Four terms are given to explain what learning in “Learning Environment 1” (their name for their Learning Operating System) will look like. Will L.E. 2.0 be next? When they move into VR headsets will it be L.E. 3.0? Why does this all sound familiar? Anyways, these four terms are: Interconnected, Personalized, Connected With Purpose, and Ethically Driven. Current education is interconnected (a lot of it takes place online with students and lessons from around the world – my son’s classes do this all the time), connected with purpose (state level standards like the TEKS in Texas are full of community cultivation and sense of meaning requirements), and ethically driven (there are many ethics laws in place (for now at least), even though enforcement of many of them fall short of course).

Then there is personalization. It seems like education focuses a lot on personalization these days, but hardly ever do we hear people thinking in terms of “autonomy,” “self-directed,” “self-determined,” etc. These are all problematic terms in different ways, but terms like these are still more about giving power to the learner. “Personalization” is often envisioned hand in hand with a lot of top-down control. Think of those catalogs that let you send “personalized” gifts. You can put a personal name on it, choose the font, maybe say you want three cookies instead of two… so each gift is “personal” and “unique.” But the company still controls what you can give, where the name goes, what comes with the set, etc. Personalization still constrains what happens – in the case of LE1, it will be the programmers and instructional designers that put together all of the options.

Which isn’t too different from how schools currently operate. Students with special needs get personalized 504 plans. Students learning to read get a custom book to read after they finish the last one (I have volunteered at my son’s school enough to even get to help clean up the entire room they have of personalized learning booklets). Personalized lessons online will assess students understanding of a given topic (Math is a popular choice) and give students different lessons depending on how well they did the last one. IXL is one such program that my son has had to work with (their website and advertising haven’t changed much in 10 years, FYI). All of these options can also quite easily slow down and repeat things where students are struggling.

But the algorithmically driven online ones are… not designed well. By the same type of people that will probably be designing LE1. I have tried to help my son work through frustrating IXL personalized lessons so many nights. Personalization has become a curse word around our house.

Redefining What Learning Means?

The About Us section starts off with “at Matter and Space, we are redefining what it means to learn in the age of Ai.”

Do we really have a definition of what it means to learn currently? Is it going to classes? Reading books in bed? Watching YouTube videos? Taking your teen to an empty parking lot to show them the basics of driving?

Or are people currently not learning from the existing Ai tools that are already in our schools? Have we really even learned enough about how the brain works to be able to confidently say what it means to learn at all?

“This isn’t just another skills platform or Ai tutor it’s a revolutionary approach to human development. By combining cutting-edge Ai with a holistic focus on personal growth, we’re creating an entirely new way for people to learn, evolve, and thrive.

With all of the safety and accuracy and bias concerns leveled at Ai – is it really wise to connect students to untested cutting edge versions of it? This isn’t the first company to focus on the whole person, and not even the first Ai company to focus on the whole person. “Entirely” new? Can they say that when nothing on their website is really that new?

“For too long, traditional education has left many behind. We’re changing that. Our platform is designed for those who need a different path—one that integrates lifelong learning with well-being, equipping individuals not just to advance in their careers but to flourish and grow as professionals and human beings.”

There it is: poor, pathetic “traditional education.” Always leaving people behind. Certainly this is not because of political powers bent on destroying public education and independent thought. Underfunding and the forced bad pedagogy of state assessments certainly aren’t part of that… right? Certainly these external pressures and problems will go away with the right tech – no need to solve systemic issues first.

Hopefully the current system of “traditional education” isn’t giving most businesses exactly what they want in employees. Because it sure seems like the traditional system is not leaving everyone behind, and they are getting hired in large numbers. Because even if Matter and Space does all it sets out to – it’s totally going to create the docile, complacent workers that don’t think outside the box that so many companies are hiring these days… right?

The Technology is Here. Or Is It?

The “Butterflies” short film ends with statement “The technology is here.” I guess in a broad sense, Ai is here, heads-up display glasses are here, real-time feedback systems are here. But Ai has a long way to go to be trusted to give accurate answers without bias to children. And the glasses featured in “Butterflies” would either have to be much more obtrusive, or need to be charged every 15 minutes to be “here.” Maybe they are just talking about the app?

If Matter and Sound wanted to improve education now, they could have just taken the money from the film and website and bought food for kids at a school somewhere. I know that sounds harsh to say, but I don’t mean it harshly. It’s just the reality that meeting a real need today would be more helpful to education over the next 5-10 years than this project will be in the same time frame. Because we are at least 5-10 years away from any of what they propose being a possibility.

The reality is that the technology just isn’t here – not at the level needed to “personalize” like their videos imply. You would have to pay a large number of experts to curate feedback content to insure useful information is presented in the correct context. “Why not just pay people to teach directly?” one might ask.

Even if they can achieve the technology – how would you finance placing it in schools (and maintaining it in working order) in a current political climate hell-bent on defunding education and tightly controlling what students learn to make sure misinformation is taken as facts?

LeBlanc’s comments about overcrowding are concerning and problematic – putting better tech into a crowded classroom is not going to solve the problems caused by overcrowding or improve learning. It will just cause more issues. Why not work on the problems of overcrowding first? And what about the kids without food to eat? What good is a personal Ai tool going to be to someone that is starving? The hopeful tone he uses when glazing over the problem of overcrowding is chilling.

And does it cause anyone at the project to pause when they hear that the current U.S. government is considering writing orders to incorporate Ai personalized education into all education as a form of control? When you are on the same side as people fighting against “DEI” and “wokeness” – shouldn’t that cause you at least a few moral dilemmas?

And let’s say that they can leap frog the tech faster and overcome political and societal issues and LE1 works as promised… who is going to hire these learners? Yes, education has value beyond just job skills, but their website does mention helping with job skills. Traditional education has left many behind not because it wants to, but because so many employers don’t want those that are left behind. Those learners are left behind by design of certain employers manipulating the hiring pools to create specific types of needs for specific people, which in turn has led schools to focus on producing those kinds of graduates.

Of course, it is unclear who Matter and Space is focusing on. They mention those left behind by traditional education. But other times they mention getting their tech to every student in a classroom. Not all students are left behind. What does that mean to be left behind anyways? Many students – even in poorer underfunded schools – graduate with good grades but still feel left behind. And what counts as “traditional” education? My son’s schools have all had different structures and systems from each other, and none of them looked like they did when I was in school. Other than having structured scheduled things called “classes,” very few of those classes look the same. Lumping them all together as “traditional education” seems to be very inaccurate and problematic. “Left behind” is vague and feels… like it was maybe used in place of another word?

Why Does Any of This Matter?

I will probably be accused of being a complainer or mean-spirited or whatever for writing all of this. None of that is true. I am truly shocked by what I see. Leaders of a well-funded project are saying things that happen hourly in education like slowing down and taking breaks have “never existed like that.” Was that something that was just a slip of the tongue? Was she referring to the Ai agent itself? Even that has been done before. Others are saying there is a “best pedagogical model” when it is really more complicated than that. Was that provocative hype, or a basis for an entire model? Or what about the statements about giving an Ai device to every student in a crowded classroom and thinking it will somehow succeed where earlier one-device per student initiatives have failed? There are so many things about this project that are just… concerning.

I remember sitting around at the LINK Lab with George talking about how problematic it is to claim something like there is a “best pedagogical model,” or how bad it is when some tech company goes overboard on hype, like claiming they are “creating an entirely new way for people to learn, evolve, and thrive.” What has happened since then? Some people have compared some of his statements to an Elon Musk-like provocation. I hope he realizes this is not a good thing.

My guess is that Matter and Space was what Siemens had in mind when he stated that “course design should cost zero.” He made that statement in response to a different Ai-fueled idea that others have also pointed out would not cost zero as well. I am sure there are already several posts out there proclaiming that Matter and Space proves Siemens was correct about zero course design costs. But everything on the Matter and Space website would point to a very expensive design process… albeit one that is just not confined to the traditional idea of a “course” (which hasn’t really existed in a while, so I’m not sure what I am even referring to here). I am all for learning happening outside of classrooms, as well as helping people that aren’t fitting in the current system. My whole dissertation was on this very topic – and the results indicated that what is proposed here is unlikely to really accomplish any of that. I am totally biased, but I still think self-mapped learning pathways (SMLP) are a better idea for all of this. Mainly because SMLP empowers the learner to do what Matter and Space is going to spend big money on an attempt to create an Ai system to do. One of the ideas we floated at the LINK Lab was to use Ai (the kind that existed in 2016) to guide learners through a custom pathway of their creation, but research into that proved unpublishable because students were really “meh” to the idea of Ai guidance (even when designed well). So, sure, write me off as a bitter ex-academic or whatever. But my solution for this issue would just cost the salary of a good instructional designer to create a pathways course.