About a third of high school teachers have used artificial intelligence tools to help with a time-consuming task: Writing letters of recommendation for students.
That finding comes from a report released recently by foundry10, an education research organization.
It’s easy to see why some teachers might turn to ChatGPT and similar tools to write recommendation letters. Nearly 40 percent of teachers who responded to an Education Week LinkedIn survey in 2022 said they write more than 10 letters of recommendation every year. And 10 percent said they write more than 30.
“I’ve used it to start a template for recommendation letters. I’ve used it to write parent letters. So I think it’s really helpful,” said Melissa Millington, a high school counselor in Missouri. “I’m honest with my students about that. I tell them, if you’re going to use it, because I know they are, here’s the ethical and appropriate way to do it.”
Each of recommendation letter can take up to an hour or longer to complete, some teachers have estimated. That means some teachers could spend up to 30 hours or more knocking out recommendations. To top it off, the task may not be part of teachers’ union contracts.
“It’s so much additional work outside of your already packed school day,” said Jennifer Rubin, a senior researcher at foundry10 and the lead author of the report. The survey included responses from over 400 teachers.
About half of the 130 teachers who said they used AI to help with their recommendation letters did so to take the stress out of the task, according to foundry10’s survey. And about a third of the 130 said they believed AI tools improved the quality of their letters.
Far more teachers, 267 in all, said they did not use AI tools to help craft their letters of recommendation. About half said they steered clear of AI because they worried that their letters would lack a personal touch. And a little over 40 percent worried about the ethics of passing off work written at least partially by AI as their own.
Could teacher use of AI hurt a student’s chances of college admission?
Rubin hasn’t studied whether teachers who use AI to write their recommendation letters are ultimately hindering their students’ chances of getting admitted to a college, securing a scholarship, or winning an award.
But that is something she’s wondered about. She thinks the answer may depend on whether the teachers use the tool for brainstorming and proofreading, or to write the entire essay.
“What comes out of ChatGPT can be generic,” Rubin said. “And often, these letters of recommendation are about personalization, and really making sure to have a student stand out.”
Using the tool for polishing and idea generation may be “quite different than having it write a letter of recommendation and not really changing much,” she said.
What’s more, teachers thought students who used AI to write their college essays acted less ethically than teachers who used AI to craft recommendation letters.
In Rubin’s interviews with teachers, they justified the difference this way. Teachers already possess the skills to write well and are fulfilling a job duty, while students are still in the process of learning how to write cogently and demonstrating those skills.
“Teachers aren’t necessarily learning new skills in the same way that students are learning skills,” she said.
Students need to show colleges, potential employers, scholarship granting organizations and others who might ask for an essay that they can write effectively.
But teachers, who aren’t trying to demonstrate their own acumen, may use AI for “rote tasks that they have to do,” Rubin said, including lesson planning, creating quizzes … and crafting recommendation letters.
