LLM-Enabled Assignments create new opportunities for students to demonstrate their ability to think … More
Initially met with mixed feelings of excitement and anxiety, generative Artificial Intelligence and large language models, such as ChatGPT, have posed a significant challenge for educators everywhere: How can we effectively incorporate powerful AI tools into learning without compromising educational integrity? As AI tools become more accessible, this question has grown increasingly urgent.
Today, Instructure, the creators of Canvas Learning Management System, and OpenAI, the artificial intelligence research organization and developer of ChatGPT, have joined forces to present a compelling solution. Their innovative partnership shows how AI can become a normal, helpful part of everyday educational experiences, greatly improving teaching and learning processes.
Instructure CEO Steve Daly succinctly summarizes the strategic vision in his product announcement: “We’re committed to delivering next-generation LMS technologies designed with an open ecosystem that empowers educators and learners to adapt and thrive in a rapidly changing world.”
This strategic collaboration embeds OpenAI’s powerful generative AI capabilities directly into Canvas, marking a new era where AI is actively utilized in the learning management system to enhance student learning, rather than being hidden as an optional tool that students might or might not use. Central to this approach is a commitment to keeping educators in control, ensuring that AI tools support and enhance human teaching rather than replace it.
AI An An Embedded, Seamless Part Of Learning
The core philosophy guiding this integration is that AI should not merely be an external tool but a seamlessly embedded part of the learning environment. In the same way that the learning management system relieves instructors of many of the administrative tasks associated with classroom management, the AI agent becomes an always-available teaching assistant, ready to support the instructor. As part of this initiative, Canvas introduces functionalities such as automated translations, rubric generation, and sophisticated grading support directly within its familiar interface. These tools aim to significantly enhance workflow, enabling educators to spend more time on meaningful interactions with students.
LLM-Enabled Assignments: Reinventing The Educational Conversation
A cornerstone of Instructure’s new AI-driven initiative is the introduction of LLM-enabled assignments. Shiren Vijiasingam, Chief Product Officer at Instructure, emphasizes this innovative feature: “This new type of assignment allows students to have rich, casual, and interactive conversations in a ChatGPT-like environment they are already used to using. In that process, they create visible learning evidence that teachers can confidently use, as it’s mapped to the learning objectives, rubrics, and skills defined by the teacher.”
These conversational assignments encourage students to actively engage with AI-driven personas, addressing complex questions and problems through dynamic dialogue. Instead of passively receiving information, students explore concepts, test ideas, and improve their understanding through iterative, responsive interactions.
Vijiasingam provides an example of a course on economic policy. Within Canvas, instructors can create assignments where students are instructed to take a position, e.g., pro or con on government intervention in the economy, and the LLM is prompted to play the role of John Maynard Keynes. The student can argue the position with the LLM, making points and responding to critique, while the dialogue is being captured and evaluated. When the student is finished, they indicate that they are ready to submit their work. Afterward, the LLM can assess the work using the course rubrics and enter the results into the gradebook, along with suitable feedback for the student. This can be sent automatically and saved as a draft for human review and improvement before it is finalized.
What is most significant about this type of assignment, compared to a static answer to a question, is the insight it can provide into the learning process rather than just the final work. Paying attention to the process is a key part of the most effective types of instruction, but it has traditionally been impossible to do in a classroom setting. The LLM allows each student to have an experience more similar to working one-on-one with an expert tutor to explore ideas, rather than sitting alone in a room writing an answer that may or may not be read, and that may or may not receive feedback, which even if it is read, will come long after the thought process has occurred.
IgniteAI Agent: Automating Routine Tasks To Enhance Instruction
A particularly exciting advancement in this AI-enhanced LMS is the IgniteAI Agent. Acting as a “conductor” within Canvas, IgniteAI streamlines numerous administrative tasks that are typically viewed as time-consuming barriers to instructional creativity. One example highlighted by Instructure’s team involves easily extending assignment deadlines for specific student groups, such as athletes, without manually adjusting each assignment. Educators can simply state their needs, and IgniteAI promptly and accurately carries out the necessary actions.
“When you can just say, ‘This is what I want to do,’ and have it do exactly that,” explains Melissa Loble, Chief Academic Officer at Instructure, “it’s much nicer to have an agent just go do those actions.”
Another example of Canvas’s AI integration is a tool called “Discussion Insights,” which offers real-time assessment of student contributions based on instructor-defined rubrics. With visual feedback clearly indicating student performance areas as green, yellow, or red, instructors can quickly identify areas for improvement, allowing for prompt and targeted interventions.
Within Canvas, the AI enables iterative and genuine assessment, allowing educators to review multiple drafts of student work with less effort. This iterative process not only provides richer feedback to students but also fosters a more authentic learning experience that aligns with real-world expectations for how work is created and assessed.
Ever Patient And Ever Available
One of the benefits of AI for learners is that it is always on. This can be important when students find themselves with pressing questions during late-night sessions. Rather than sending a question by email that may not be answered until the next day, students can dive into a session with the AI. Leah Belsky, General Manager and VP of Education at OpenAI, remarks, “Students like using AI because it’s an ever-patient tutor. No question is beneath it, no interaction is too long.”
One promising future development is the integration of AI with digital twin technology, which involves AI-powered simulations of real-world instructors, such as those provided by companies like Praxis or Tavus. With a digital twin of the instructor, students may feel more comfortable taking risks or trying out wild ideas without fearing academic penalties. It’s often said that failure and risk-taking are vital parts of learning, but this is hard to practice when the only chance to take risks is during final assignments. The AI-enabled environment allows learners to step onto thin ice, knowing they can recover and adjust before the final assessment.
Similarly, Loble emphasizes that “using AI in classrooms allows all students to participate in the conversation, not just the ones who raise their hand and have their voices heard.” This ability to involve everyone is especially important in seminars, where 20% of the students often generate 80% of the comments.
Institutional Benefits And Accreditation Support
Canvas is also looking ahead at using AI to support broader institutional goals, including accreditation processes. AI-driven features, such as “Ask Your Data” and “Intelligent Insights,” can proactively highlight potential instructional issues, enabling institutions to manage faculty performance and student engagement effectively. Such tools can significantly reduce administrative overhead by capturing information needed for these key processes as it is generated, rather than as it is needed for submission.
Lifelong Learning And Career Integration
Instructure is expanding its AI-driven educational innovation to the workforce with its newly announced “Canvas Career.” This platform emphasizes skills-focused, personalized learning paths and professional development, showcasing AI’s ability to support lifelong, adaptive learning. AI-powered recommendations and adaptive learning experiences aim to seamlessly connect traditional education with professional growth.
Looking Forward
The Instructure-OpenAI collaboration signals a robust, ongoing commitment to leveraging AI to genuinely enhance educational experiences. This groundbreaking partnership lays a strong foundation, but as the teams behind the initiative note in their announcement, it is “the first of more.”
As AI is systematically integrated into the learning management system, and as more and more data on its effective use becomes available, the flywheel of innovation is sure to spin ever faster.
