Generative AI products such as OpenAI’s ChatGPT, Google’s Bard, and Anthropic’s Claude have shown tremendous potential as tools that can support students in the learning process. This is particularly true of learning that is language-intensive or process-driven. While students may dream of tools that can free up their time by doing their assignments for them, the true potential for these tools is to make it possible for students to do more work and to make that work more effective by ensuring that students do the majority of their work in their zone of proximal development. While such fine-tuning of assignments has always been possible when teachers are working 1-1 with students or in small groups, it has been challenging, if not impossible, when working with entire classes of students, especially when those students are of varying degrees of ability and achievement.
Enter Class Companion
Class Companion, a new AI platform that harnesses the potential of generative AI to provide high-quality feedback to students, publicly launched today, announcing the receipt of $4M in seed funding round led by Index Ventures. Class Companion, free for teachers to use, has grown from a handful of teachers when it first became available six months ago to now being used by teachers at over 5,000 schools.
Teachers have been the focus of Class Companion’s mission from the outset. “If we want to use AI to improve education, we need more teachers at the table,” said Avery Pan, Class Companion co-founder and CEO and a former writing tutor.
With Class Companion, Teachers can now provide writing assignments to students secure in the knowledge that students will be able to get immediate feedback on their writing, at the moment that it can be most helpful, rather than having them do the work and then have to wait until the next day, or next week, to receive substantive comments.
Timely Feedback is Essential
The ability of the tool to provide immediate feedback is where its true promise resides. Feedback is essential for student learning. But for feedback to be effective, it has to be delivered at a moment where it can be understood and acted on by the student—two attributes that traditional home often lacks.
Traditional homework can be likened to a student shooting free throws but not being told if the ball is going into the basket until the next day. No matter how good the comments are the next day, the student will not be able to make the changes needed to improve without being in the headspace they were in when doing the work.
The importance of immediate feedback becomes more evident when considering a student who sits down to complete an assignment when fundamentally confused. Such a student will wind up either spinning their wheels, not knowing what to do, or they will do the wrong thing, iterating on their confusion, making it much harder to address in the future.
Errors as Opportunities
Another challenge about traditional homework is the tension between seeing errors as opportunities for learning and improvement and seeing errors as reducing a grade and, for the catastrophizing student, ultimately the likelihood of gaining admission into one's desired college.
For errors to become tools of growth rather than flaws to be hidden, the ability for students to endlessly iterate their work is needed, each time being able to get feedback that can be folded into the next iteration. This opportunity shifts the focus from the product to its production process. These approaches have long been advocated for by initiatives such as Grading for Equity. Still, they can be challenging to implement, especially for a teacher with a large class and a pacing guide they are expected to follow.
This is ultimately the value of AI tools. As they are all watching, they have the potential to spot errors as they occur and can learn to respond to those errors at the exact right moments to ensure maximal learning.
Particularly noteworthy is that the Class Companion does not see the AI as the ultimate arbiter of truth or falsity. Instead, students can easily contest the critique they receive from the AI and can appeal to the teacher for a final determination. This increases trust on the part of students and will also increase self-advocacy and a willingness to spot errors in assessment. This way, feedback is transformed from absolute truths handed down from on high to the starting point for a conversation.
Providing Teachers Flexibility
With Class Companion, teachers can select from an extensive repository of assignments or add their own. No tech skills are needed in either case. The assignments cover questions from short answers to developing outlines to completing full-length student essays. In each case, students are prompted, and their responses are evaluated by the tool, which provides targeted feedback on student work.
One feature that teachers enjoy is that the feedback can be shaped to focus on specific skills - allowing customization on the class level and individualization on the student level. Teachers report that this ability to focus on areas of greatest need for students has led to them changing their approach to teaching.
Perhaps the best measure of promise for any digital tool is when students complain not about its presence but its absence. Tom Richey, a teacher in South Carolina, said, “My students really enjoy it. Last week, my students were doing an assignment on paper, and they asked, why aren’t we using Class Companion?”
This will ultimately be the test. Unless a product is easy to use and delivers outsized impact, odds are it will fall by the wayside. Class Companion is off to a promising start. I look forward to following its evolution.