Thank you for the heart, creativity, and resilience you bring to your classrooms every single day. We know the work is as challenging as it is meaningful—and because of educators like you, students are questioning, creating, and growing into people who feel more capable of shaping the world around them.

This Teacher Appreciation Week, we’re celebrating our favorite people (teachers!!) all week. Today, we’re giving away a teacher survival kit on Instagram, and supporting classroom projects requests on DonorsChoose 💛We hope you’ll join us!

Join the amazing community of teachers using Figma’s powerful (and free) tools in their classrooms.

Here are three busy discussion threads in our DTT Community right now …

You can join the free DTT Community to read ideas like this, connect with other educators, discuss, and share ideas of your own.

Searching for that a fresh activity that connects kids to the content without requiring a mountain of prep work. If you're looking for a "low-prep, high-impact" strategy you can implement tomorrow, we have you covered with an oldie but a goodie… "Caption This!".

Using Google Drawings or Slides, this activity turns simple images into powerhouses for critical thinking. By adding a speech or thought bubble, you let students "become" the subject of a photo, forcing them to make inferences and understand different viewpoints.

Our amazing DTT community has taken this template and run with it! Here are some of our favorite ways educators are using it in the wild:

Check out how these educators from our Wakelet collection are using it across the curriculum:

When ChatGPT was introduced in 2022 (and generative AI subsequently mushroomed), I totally understand why we wanted AI detectors (aka “AI checkers”).

The problem? The technology works very differently between the two types of detectors.

The bottom line: Lots of teachers are still using AI checkers to “catch students in the act of AI cheating.” In reality, the checkers are just guessing based on common patterns, so this method is really inaccurate.

There are a few things that AI checkers measure to guess whether text was AI-generated or human-generated:

When it’s done, it’ll provide a score — often a percentage — that usually states what it thinks the probability is that the text is AI generated. Think of it like a confidence score … how confident the AI checker is that the whole text (or certain passages) are AI-generated.

If you go to the AI checkers’ websites, they’ll tell you how incredibly accurate they are. But there are ways to manipulate that data. (Or they could just flat-out lie.)

Here are just a few of the findings …

This is just scratching the surface.

After looking at the research, institutions like Vanderbilt University disabled tools like TurnItIn’s AI detector. In its guidance to university educators, it stated:

Five things that give me pause about AI checkers …

AI models are trained on human-created writing — AI models like ChatGPT use lots of human-written texts to learn how to write. They are intended to sound very human. When humans write text — and when AI writes text that’s supposed to sound human — the margin of error becomes tiny. It’s a LOT to ask an AI checker to determine that with certainty.

It’s all based on probability — That means guessing. AI checkers are looking for patterns. Sometimes, your students’ writing has the same patterns as AI.

We’re sounding more like AI these days — The more that AI-generated writing is circulated commonly in the world, the more likely our human speech and writing patterns will start matching it. We start mimicking what we see — and, again, the margin of error becomes even tinier.

It’s based on a narrow-minded premise — AI checkers exist to answer the question “Did you use AI?” The more we use AI, the more we realize that it can be used in a variety of ways — some helpful, some unhelpful.

As time goes on, it will only get harder — As Sadasivan et al (2023) showed in a study, as AI models improve and produce more human-like text, the gap between AI-generated and human-written will get smaller and smaller … until eventually, it’s almost like random chance (a flip of a coin).

In a recent thread I posted — “VIDEO: About those AI checkers,” here’s some of the discussion our DITCHers (aka community members) had …

(Not a member of the DTT Community yet? Click here to join for free!)

I wish I had a quick “do this one thing” answer for this.

The reality, I believe, is that it’s going to take lots of small shifts — shifts that will work in some instances but not in others. We’ll need to adjust assignments, help students to be aware, make ourselves more aware, and find different ways of navigating all of this.

Here are some thoughts …

Tell students why we’re doing this assignment — Sometimes, students don’t understand how they’ll improve as a result of doing our work. If they understand the “why” behind it, they might be less likely to farm it out to AI.

Provide the right amount of scaffolding — Sometimes, students don’t see how they’ll ever complete the assignment (or don’t believe enough in themselves to do it). A little support could be what they need to do it themselves.

Chunk activities — Instead of asking students to do large assignments over multiple days, turn it into smaller parts (“milestones”). Break large activities into a few parts. Help them to see the steps to complete the whole.

Analyze why (and how) students misuse AI — If we put ourselves in our students’ shoes (or even talk frankly with them), we can see why they (or their peers) might use AI and identify how to support them and adjust assignments to keep learning going.

Practice persistence — Students might turn to AI when things get tough. When they realize that struggle is a part of learning … that they don’t have to be perfect … that they can do it if they stick to it, they’re practicing persistence. If they’re bad a persistence, help them to grow it slowly like a muscle.

Reconsider sending work home — I gave up on homework years ago. But as a high school Spanish teacher, I see a path to covering my content and standards without it. In some subjects, it’s almost impossible to meet goals without homework. We know that motivation is often lower at home to do work than it is at school … and students will look for ways to get out of it (or get it done as fast as possible).

Use paper instead of tech — Lots of us are considering this. There’s still a lot I do with paper in my own class. We do have to keep in mind that eliminating tech does remove important accommodations like assistive technology (screen readers, text to speech, translation, etc.).

We’re having a conversation about this in our DTT Community.

Click here to jump over to our discussion about AI checkers to read what others wrote — and to share your thoughts.

Not a member of the community yet?

There are even more ways I can support you in the important work you do in education: