Dynamic Simulations vs Standard Learning Videos: What A Parent-Teacher Learned Testing New AI Learning Tools
Last Tuesday, I found my 4th grader Javi slouched on our living room couch, a half-eaten bowl of Goldfish crackers tipping off the edge of the coffee table, a pre-recorded science video droning on the TV. His phone was tucked under his sweatshirt, and I could see the faint glow of Roblox bobbing as he scrolled under the table. He was supposed to be studying erosion for his quiz the next day, and after 15 minutes, he couldn’t tell me the difference between weathering and erosion. I’d been there a hundred times: standard educational videos work for some kids, some of the time, but when a concept just isn’t clicking, they’re just background noise. A part-time para at my kids’ elementary school, I’d just gotten an email from the tech coach about a new AI tool that builds dynamic simulations instead of using pre-made standard videos. I was skeptical — I’d tried dozens of shiny new learning tools that fizzled out after one use — but I was out of ideas, so I pulled it up on my laptop.
The tool didn’t pull up a pre-scripted video. It opened to a blank landscape, and let us tweak any variable we wanted: slope of the hill, type of soil, amount of rain per minute. I pointed to Javi’s study guide question: “What happens to erosion when a river gets steeper?” Instead of telling him the answer, the simulation let him test it. First, he messed around for a minute, built a super steep hill out of sand and cranked the rain up to maximum. The whole hillside washed into the river in three seconds, and he dropped his phone, leaned in, and laughed. That was the first time he’d looked up from his phone all homework period. He spent the next 10 minutes testing soft soil vs rock, gentle slope vs steep, light rain vs heavy. When he went back to his study guide, he answered every question without looking at his notes. He even got the tricky one about how faster water carries more sediment, which the video had stated outright but never clicked for him.
That doesn’t mean it’s all perfect, though. The next night, we tried it to work on equivalent fractions, a concept Javi’s been stuck on for weeks. The simulation uses pizza slices to show how different fractions can be equal, but when Javi tried to split a pizza into 12 slices to compare 3/12 and 1/4, the AI glitched. It counted 3 slices as 1/3 instead of 1/4, and we had to restart three times before it got it right. Another time, when I used it with a small group of 5th graders working on force and motion, one kid spent 10 minutes just launching balls off the end of the ramp into “outer space” instead of testing how height affects speed. It’s way easier to get distracted with a simulation than it is with a pre-recorded video, because you control what happens. That’s the upside and the downside.
I’ve been using it for a month now, both at home and at school, and I’ve landed on two simple rules that work for me. First, I only pull out dynamic simulations for process-based concepts — things that are about how change happens, cause and effect, movement. That covers things like erosion, force, fractions, how a bill moves through government, the water cycle. For straight memorization, like the order of the planets or state capitals, a standard 3-minute pre-recorded video works just fine, and it’s way less likely to lead to distraction. Second, I always build in 2 minutes of unstructured play first. If a kid wants to launch balls into space or flood a whole valley first, let them. That curiosity is what makes the simulation work, and if you get the messing around out of the way first, they’re more likely to focus on the actual question once you start. I’ve tried this in the classroom and at home, and it cuts down on off-task behavior by more than half. It’s a small tweak anyone can do, no fancy training required.
Last week, I worked with a student who has ADHD and almost never participates in small group science. We were doing force and motion, and he couldn’t get why friction slows objects down after watching two different standard videos. We pulled up the simulation, and he got to test a ramp covered in sand vs a ramp covered in wax. He rolled the same ball down both, and watched how far it went each time. Halfway through, he looked up at me and said, “Oh, the sand grabs the ball more, right?” That’s a moment he never would have had just watching someone else explain it. He got it, on his own, because he tested it.
I still get asked if I actually use this thing, or if it’s just another gimmick. The answer is yes, but not all the time. If I’m trying to get dinner on the table and just need Javi to watch a quick video to review for a quiz, I still pull up a standard pre-recorded one. It’s predictable, it doesn’t glitch, it doesn’t need my input. But when a kid says “I don’t get it”, when a concept just isn’t sticking after watching a video, I reach for the simulation first now.
It’s not perfect. It makes mistakes sometimes. It takes a little more of my time than just hitting play on a pre-made video. It doesn’t replace me or the kid’s teacher, it just gives them a way to test their own ideas instead of just memorizing what someone else tells them. Right now, we’re still figuring out what works and what doesn’t. Javi still aced his erosion quiz, though, and the thing he said he remembered most a week later was that big flood he made. That’s more than he can usually tell me after he watches a standard video, so I’ll keep using it. We’ll just stick to the 2-minute play rule, and keep an eye out for glitches. It’s working well enough so far.
Please indicate: Thinking In Educating » Dynamic Simulations vs Standard Learning Videos: What A Parent-Teacher Learned Testing New AI Learning Tools