As teachers head back to school, the topic of AI may or may not be weighing on their minds. Between getting their classroom ready and student roster set, thinking about the moral implications of ChatGPT might not be top of mind. But that doesn’t mean we shouldn’t spend a little bit of time addressing it.
Throughout this summer, I’ve been traveling to schools across the country to talk with teachers about the pros and cons of AI in schools. These conversations have given me great insight into the concerns of many teachers and what AI means for learning going forward.
The workshops have been designed around unpacking what AI really is, what it can and can’t do for us, and, ultimately, how it can help our students. They listed many concerns around cheating, academic integrity and a lack of critical thinking. This snapshot below was taken from a group of 100+ high school teachers I recently worked with:
If you are someone who is talking with your staff about AI, I think it’s extremely important to hit these concerns head-on. Going through this practice helps both inform and provide guidance on where your focus should lie. What follows are seven strategies that I use when sharing about AI with staff from pre-K through higher ed. While I could use a tool like ChatGPT to help me, I also want to point out that these are not AI-generated ideas. They come from working with actual teachers and schools around the use of AI in their classrooms.
1. Identify AI in everyday life
Artificial Intelligence was invented in 1951. It’s been a part of our lives far longer than the past few years, but the launch of ChatGPT in November of 2022 raised awareness of “Generative AI” throughout the world. After identifying fears around AI, I invite teachers to identify how they use AI in their daily life. Listing these out can help alleviate some fears around AI use.
The most common responses are the following:
- Digital assistants like Siri and Alexa
- GPS/Map apps that help with real-time traffic alerts
- Personalized recommendations on platforms like Netflix and Amazon
- Autonomous vehicles and driver assist functions
There are many more that come up in the follow-up discussion but it does help alleviate concerns when we see how helpful it can be when used appropriately.
2. Identify ways we use AI in schools
As a follow-up to the above discussion, I then ask for ways that they use AI in their daily teaching life. While this list isn’t as exhaustive as the ways they use it in daily life, there are lots of different ways it is being used or can be used. Some examples shared are:
- Auto-graded assessments
- Speech-to-Text and other adaptive software
- Predictive auto-fill in documents and emails
- Language programs like Duolingo
Once again, we are truly scratching the surface when you bring in generative AI, but many other AI tools have been used in schools for years.
3. Check your policies
I had an opportunity recently to host a leadership summit with Tech & Learning before the ISTE Live conference. This summit brought 60+ educational leaders into a room to discuss the current state of educational technology. When asked how many were concerned about AI in their institutions, every hand was raised. However, when I asked them how many had written it into their Acceptable Use Policy (AUP) only two hands went up.
This technology is growing at a rate that is akin to the explosion of the app store back in the early 2010s. Need a tool? There’s an AI app for that. With this rapid increase in generative AI tools, we also need to address data privacy and ethical concerns. Many of the tools have 13+ age restrictions but we also know our kids have access to devices outside of our schools.
This past month, I collaborated with All4Ed and Microsoft on a resource for schools to utilize when tackling AI. This Future Ready Emerging Practice Guide is a great place to start for a school leader struggling to wrestle the AI beast. The sooner you start this conversation, the better.
4. Take time to explore
ChatGPT definitely spearheaded the current generative AI movement, but, as I mentioned above, there are hundreds of other tools out there with some component of generative AI. When exploring with teachers, I break AI into these 6 main buckets:
- Prompt Generators (like ChatGPT, Bard, etc)
- Image Generators (like Canva Text-to-image)
- Slideshow Generators
- Quiz Generators
- Cheat Detectors
- Creative Tools
While I think it’s important to cover the prompt-generating tools, the other categories I let them dictate what they want to learn more about. I think it’s also important to give teachers time to try out tools under each bucket. For example, I might give them 10-15 minutes to try out a series of prompts using ChatGPT, Bard, and Canva’s Magic Write. Then I’ll ask them to share what they discovered.
The important part of this step is letting teachers explore. I recently created this post on 10 creative AI tools as a way to get started, but I know these tools change almost by the minute.
5. AI is not perfect
Like any technology tool, there are times it works well and times it doesn’t. AI is no different. It is prone to “hallucinate” information that may or may not be true. In asking teachers to explore prompt generator tools, most responded with the same thought: It’s a good first draft.
I recently interviewed a good friend and USC professor of public diplomacy on my Forward to Different podcast. Rather than forbid his students from using it, he required them to use ChatGPT to build their diplomacy argument. He said that on the surface, it produced a B-minus quality paper. He then told the students their job was to make it an A-level paper. In order to do this, they would have to know their material and resources to a high degree to improve the AI output. He shared that the quality and knowledge they demonstrated as a result was notably more thorough than previous semesters.
Knowing that AI can provide a “good start” for students trying to synthesize or analyze a concept seemed to put many educators at ease. Much like when I taught elementary school, I could tell when a student’s parent did the assignment for them, AI is no different. That said, as cheating was the number one concern with AI in schools, it’s important to address that head-on as well. Hence the next bullet point:
6. Creating “AI-resistant” assignments
The ethics around using AI to help with an assignment are still largely ambiguous. Many teachers feel like using a tool like Bard to help you write a paper is cheating, while others feel like it could be helpful for learning. A young student struggling with writing could generate a children’s story using a tool like StorySpark. But could tools like these rob students of learning foundational skills?
It’s a fair point, but I also think it is wise for those of us in education to consider that maybe learning AI will also soon be a foundational skill. We tend to focus most of our energy in education on the product rather than the process and AI exposes that over-reliance. Evaluating the process of a student’s assignment is impossible for AI to replicate. Things like your own thoughts, reflection, and using some sort of hands-on or verbal component in the assignment are key.
The University of Michigan recently published this guide to avoid AI-based cheating. It’s a solid guide with many great points but if you scroll all the way to the bottom you’ll notice the following text:
Now there are “cheat detector” tools out there like NoRedInk, Turnitin, and ZeroGPT. However, like generative AI, they too have hallucinations and false positives, so use them with caution.
7. Start small with time savers
With the amount of new generative AI tools on the market, it’s easy to become overwhelmed. I mentioned in bullet 4 how important it is to give teachers time to explore and play to truly judge its usefulness. Rather than try and learn every tool, educators should focus on one or two that might be useful for them.
Prompt and chat generators can be instant time savers when for low-hanging fruit like writing a parent email, creating some stations for a particular skill, or even ideas for differentiating existing lessons for students with learning disabilities.
Educators face such a time famine in their daily work. AI could help give back some of that time and potentially allow students to dive deeper into the subject matter. Stepping out of my own comfort zone, I’ve started to design several “bite-sized” tutorials called “Hooked On AI” using TikTok, Instagram and the platform formerly known as Twitter. Here’s a sample:
Like most educators, I’m taking time to explore and determine what of this new AI revolution will not only be effective for students but also help me save time in my day. I’m excited to see where AI will take education in the future and hope that all educators get a chance to explore it this year.
