Artificially Intelligent - Part II

If you can just ask someone or (something) else, do you need to know anything?

This is the second essay in a series of four where I discuss the implications of AI in American public schools. You can find Part I here.

---

It wasn’t surprising to me when I was working at very low-socioeconomic strata schools for the first 13 years of my career that the 1st- and 2nd-generation American students making up the majority of students there didn’t know much about mythology or fairy tales or history when I was teaching them literature and, later, history. In response, I slowed my teaching down and filled them in, telling them the stories I thought they’d missed, either in school or by dint of their parents’ lack of formal education. I wasn’t surprised to see seniors struggle to do the required math to make sense of supply and demand curves or calculate elasticity of demand. I did my best to catch them up so that they could, eventually, do those things.

I left that school shortly after the birth of my second child and moved to a new high school closer to my own home. It carried the “California Distinguished School” label by dint of its student test scores and was populated by children of high-income-earning professionals and entrepreneurs in a safe, affluent neighborhood. The middle schools which matriculated students into my high school were California Distinguished Schools too.

So color me shocked when I realized their knowledge wasn’t a whole hell of a lot deeper than the kids I’d taught in the poorest schools. I could barely wrap my head around it. How was this possible?

And then I realized what it was: a paucity of cultural literacy. Even the rich kids didn’t know the stories of their culture. They knew George Washington was our first President (and a slave-owner), but almost nothing about his childhood, his stint in the British Army, or his Presidency. They knew Thomas Jefferson had been a President (and a slave-owner) but couldn’t reliably trace the Declaration of Independence to him. They knew nothing about Thomas Paine, the Adamses, Great Britain or how our government worked. The Civil War? I mean, most of them knew it was between the states (and about slavery). They knew Lincoln’s name and that he’d been assassinated, but not much more about him or the war itself. If they read books on their own, they knew fairy tales and some mythology (thanks to the Percy Jackson series), but the rest of them? Nothing.

They definitely knew that mitochondria are the powerhouse of the cell, but beyond that, their knowledge of basic biology was limited. Astronomy? Nil. Earth science? Well, they understood how earthquakes happen, but we’re in California. Weather? Currents? Basic physics? Nope, nope, nope. They couldn’t reliably name all the oceans or all the continents, let alone the countries in Europe, Asia, or Africa. It was news to them that Austria wasn’t a misspelling of Australia when it appeared in print. WWI and WWII were often and easily confused.

Please keep in mind that I’m talking about seniors in an elective Advanced Placement economics course, not the general student population. Unsurprisingly, the students in my non-honors 9th grade english classes knew far less.

Most were very familiar with popular stories on social media, and I found it very curious that they all had strong opinions on the controversies of the day, whether it be George Floyd, the value of social justice activism, proper policing, etc. Scratch the surface though, and they struggled to tell you why they held such beliefs. It was easy enough to figure out, if you listened to the words they used in conversation about these topics. They had heard messaging, usually from social media (or from their friends…who had heard it from social media) and sometimes, from their parents, but they could muster few facts to support their opinions. In effect, it became obvious that algorithms were assigning my students their opinions.

I’d sometimes hear students arguing with one another, but their arguments lacked evidence, lacked reasoning. There were lots of logical leaps over canyons of empty space. Anyone willing to stop and ask them simply, “How do you know that?” could yank them back down to earth, but their friend groups were very often built on shared opinions about the issues of the day. Their arguments were emotional, which mostly consisting of spluttered, half-remembered phrasing courtesy of a half-forgotten video.

These are all jokes now. Memeworthy, in fact. There’s a multitude of very concerning videos where a young man or woman espousing a particular belief gets the most gentle of pushback, such as a request to define a term, and falls apart entirely.

And that’s why I’m going to refer you back to Efrat Furst’s work.

See, an argument, an idea, needs to be based on that bottom right picture. There, pieces of information can be culled from all of those triangles individually, and from larger ideas, represented by pyramidal structures, as well. In that zone of learning, arguments are made from connected bits of data: factual knowledge, stories, and a person’s own experience.

Meanwhile, my students were mostly living in that top right picture, where the knowledge is jumbled, unstable, and disconnected. And this is why AI is so unbelievably dangerous: it prevents most new orange blocks from descending.

Wait… what? You’re claiming AI prevents kids from learning anything?

Yeah. I am.

This is from a previous post I wrote explaining why kids rarely write essays in school anymore:

Enter history. While I’m aware this is up for debate as it pertains to current practice in U.S. schools, history taught in junior high and high schools centers around facts. It is chronological. The names students hear in class are names they have heard in the past. McKinley the mountain becomes McKinley the man. The $1 Sacajawea coin becomes a remarkable 15-year-old woman. Their teddy bears point them to Teddy Roosevelt. The name New Mexico suddenly makes total sense. Everything that happens can be pointed out on a map. As soon as you begin any discussion of the 20th century, photographs of real life abound. Film becomes available. Students recognize skylines and monuments and natural landmarks. America is so young, people alive in the 19th and early 20th centuries don’t seem all that different to us.

In other words, in a U.S. History class students bring certain bits of foundational knowledge that allow them to hold tight to new information. Neuroscientists sometimes refer to these saved chunks of knowledge as neural hooks. This made U.S. History way stickier, learning-wise, than English.

Back in 2011, when iPhones were suddenly in even my poorest students’ hands, the spurious logic floating around the California school system (which is still there, like an evil but moronic poltergeist) was that we should focus on projects and deeper learning rather than facts and knowledge because these kids suddenly had supercomputers in their hand that would allow them to look up any fact at any time. Why bother doing all the work of teaching and ask students to do all the work necessary to turn an orange triangle into a dark gray one when Siri could just tell you whatever you needed to know the second you needed to know it? Why work hard to organize a presentation into a clear structure, come up with student-appropriate opportunities to practice, and plan out how often and how spaced the repetition should be? Why spend all that time grading and correcting their misunderstandings and misapplications? Why do all that just to turn a little incidental fact from an orange triangle to light gray to dark gray so that it could become part of a foundation for building additional knowledge? Why do that when kids can just look it all up?

They no longer need to remember.

That sentence should send chills down your spine if you understand Furst’s model.

I bet you can see where I’m going with this, but in Part III, I’ll explain how we got here, and after that, in Part IV, how I expect AI will be misused in schools, and what you, parents, can do about it.

---

If you appreciate this and believe that my essays, podcasts and lesson plans will be useful to American families recovering control of their child’s education (even if they can’t fully control their schooling) please subscribe to support my work. If you can’t afford to help, know that I intend to provide the most important posts to motivate and support you in teaching your own children free of charge.