I’ve been an educator now for over 25 years and part of that time I was a part of the initial wave of Tech Integration Specialists. Teachers with a particular interest in where things were headed were put on “special assignment” for part of their contracted hours to learn and teach digital applications to their colleagues for the improvement of the learner experience. They were supposed to make us more efficient and help kids learn better. Sounded great at the time. Because I primarily work with learners with various challenges and those from marginalized communities, one thing was always abundantly clear…virtually no tools created in this space are inclusive by default. There always has to be some kind of work around, adjustment or supplement to make the process work for various learning populations.
I once went out to support a colleague who was working with a student who was having some difficulties with a school district’s established workflow for students. Super cutting-edge stuff. Teachers were to convert all of their materials to PDF files (scanned pages with no OCR), upload them to their LMS so that students could access them for download and markup with Notability on their shiny new iPads.
My colleague was a Blind/Visually Impaired Specialist working with a blind 7th grade learner. Virtually no aspect of this cool new way of doing things worked for this student and she was spending roughly 30 hours a week trying to modify and adapt these learning materials that were hopelessly broken from the start. Millions of dollars went into this 21st century workflow and none of it worked for students who needed something accessible.
I think about all of the things I learned during that experience and others like it over the years. I spend years as part of a school district’s committee on Equity and Inclusion as part of our strategic plan, which also rarely touched on accessibility. As we reviewed district policies and procedures to “remove systemic inequities and identify our role in perpetuating them,” I became acutely aware as to how much is broken from the start.
Which brings me to this…
While I have been speaking to educators around the country, spending hundreds, if not thousands of hours by now interacting with various AI platforms, teaching online in an asynchronous setting, reading research, writing books and curriculum, developing trainings and more, I continue to develop new insights into how these tools work. And already, so many of our practices around AI in education are not only broken, but they are also creating all new systemic inequities that didn’t exist a few years ago. They may be setting us back decades.
Here’s what I mean…
The first instinct of many school districts around the country was to use internet filtering to block students (and many teachers) from even accessing AI platforms.
Inequity 1
The number of these tools grew exponentially, so there was no way to keep up with the list of sites that needed to be blocked. And what were schools going to do with those who:
- Accessed from their phone’s data plans
- Knew how to get around filters
- Used personal devices and WIFI at home
- Saw AI integrated via an API into so many of the other applications that couldn’t be blocked
The users in those categories now had access to powerful tools that the rule followers did not, and this inequity started to grow.
Inequity 2
As has been well documented, the tools themselves are biased based on their own training data which compels them to look for probability and predictability so they can sound like us, and what are we if not imperfect, biased beings?
Educators using AI to develop curriculum (without sufficient prompting) were developing content that did not suggest examples that included people of color or women. Image generators rarely demonstrated inclusive or representative faces and skin tones for even simple requests. At worst, they were demonstrating racial and gender bias when it came to images to represent various career pathways. This problem persists and with the large number of educators now using it in this way, the problem is only getting worse.
I experience this myself in a number of instances while working with these tools and asking for lists of examples or images. It’s not that you can’t prompt for something better, but if it’s not part of your core principles to seek out inequities, many in dominant cultural groups will not even notice anything is missing.
Inequity 3
There is such a wide gap right now in terms of educators who want nothing to do with these tools, those who explore them a little bit, and those who accept the inevitability and are finding ways to use them to support learners and take on some of their own massive workload. We are spread out across a wide continuum of usage that learners around the country, or even groups of students within individual schools are having wildly different experiences. This is a big deal when it comes to something this disruptive that is being quickly integrated into so many aspects of our lives. We must find ways to support our learners who will live with a quickly changing landscape at faster and faster rates than we’ve ever seen before. When it comes to technological advancement in this era, the more there is, the faster there is more (according to this new phrase I just coined.) We’re growing exponentially. It’s no time to look back to traditional methods and think that somehow stagnation is part of the solution. There’s no time for that.
Inequity 4
This one has to do with agency. Those working in areas of equity and inclusion have known for a very long time that not all people in this country have the same platform. Right now, there is legislation in some states that bans the teaching of diverse perspectives in district approved curriculum. When it comes to AI, we’ve already given agency to machines in ways that all people in this country still haven’t experienced. News agencies are using AI to write stories, bloggers are using it, marketers are using it, social media bots use it so much that humans have been starting to sound more and more like machines since 2018!
That is by no means an exhaustive list, but the more obvious and blatant ones. As districts around the country update their Acceptable Use Policies and teachers update their syllabi in time for the coming school year, do we really think that the consequences for violating these policies will be equitably applied to all learners? We know they won’t be!
These inequities, as they build, will not just sort themselves out and go away on their own. We are responsible for identifying our roles in perpetuating them and finding ways to dismantle them. We have to share our experiences with each other, collaborate on solutions and discuss what is working, what is not working, and what still seems impossible.
“…the best way out is always through.” – Robert Frost
For more on my work on AI go to Inclusive AI Strategies!