Q&A with Nicolaas Matthijs, Vice President of Product Management, Anthology
Why is there more interest and discussion about the use of AI in higher education today?
Artificial intelligence has been around for a long time in different forms, but there are some new developments today that make AI more applicable and interesting for education.
Most importantly, the AI technology of the past 20 years has relied on having access to substantial amounts of very structured data. In education, however, we usually deal with a large amount of sparse and unstructured data, as well as a significant amount of natural language. Generative AI has been designed to work with natural language and can use that information to generate new content such as text, images, or other media. This is why generative AI is so relevant in the context of education.
This is an exciting but cautionary time for us. Personally, I believe generative AI is here to stay, and there are numerous positive ways that these technologies can be used to improve the lives of students and instructors in higher education. However, we also must tread carefully and ensure that we understand the strengths and weaknesses and that we use them responsibly.
What are some of the ways generative AI could be applied in higher education?
When it comes to instructors, we see immense potential in utilizing these AI tools within the context of course and content design. At Anthology, we are focused on offering AI tools that aid and inspire instructors and instructional designers. We believe that these tools should ideally assist and support the instructor, with the goal of improving efficiency and providing a meaningful starting point. In fact, we recently received feedback from an instructor who shared how much time she saved by using the tool to generate a robust question bank; she had about 10 questions and wanted 40. The tool enabled her to get a head start.
While there is a lot of focus and discussion in the industry on using AI for things like AI detection and automated grading, we recognize those use cases carry a good amount of risk for inaccuracy and bias. We have seen examples where students are being incorrectly flagged as having used generative AI tools to do their homework. This has a profound negative impact on those students and may have a negative reputational effect on the institution. That is an example of the risk when no human is present to act as a check on AI, which we believe is critical.
Could you provide an overview of Blackboard Learn’s new AI Design Assistant, including its capabilities, and share the rationale behind its development?
Anthology’s new AI Design Assistant for Blackboard Learn is seamlessly embedded into our learning management system (LMS). We are the first major LMS vendor to bring these generative AI capabilities into the learning management system space with the goal of assisting instructors and instructional designers in creating courses faster.
We recognize both the immense potential and the associated risks of AI. This is why it is important to highlight how we initially focused our AI efforts on course and content design—a domain that empowers instructors while preserving their control over the process.
The AI Design Assistant for Blackboard Learn helps by taking the syllabus or desired learning outcomes of a course then uses that information to suggest a structure for the course complete with titles, descriptions, and visuals to provide the instructor with a starting point to review, react to, and adapt as they see fit. This can save them a significant amount of time, and we are already seeing evidence of this in our early instructor feedback conversations.
Similarly, the tool can also generate potential test questions or grading rubrics based on course content to provide instructors with a starting point. The instructor can then take over and control what they want to include, exclude, or expand from that point. This is a good example of how AI can help people rather than replace them, as some fear will happen.
How do you see AI personalizing learning for students?
Generative AI holds immense potential in revolutionizing the creation of highly personalized and adaptive learning experiences for students. However, a key principle to fundamental value of AI is the ongoing presence of human oversight, ensuring responsible use.
Consider the role of AI in helping align all course elements such as content, assignments, questions to the skills, and learning objectives of the course. Once properly labeled, an AI tool can elevate instruction by customizing content delivery to match each student’s specific areas of weakness and strength, thereby crafting an exceptionally personalized and rewarding learning experience for students.
How can college and university leaders help their professors or instructors better understand the capabilities of AI and the potential uses you have described?
This is an ongoing part of the conversation around AI; not just exploring its potential uses but asking how we can introduce these tools in the right way. Many users will need some professional development or training and will need to develop familiarity with these tools over time.
This issue informed our approach to introducing the AI Design Assistant for Blackboard Learn. That is why it is embedded in the LMS, and seamlessly integrated into what people are already using. That makes it very natural for the user. However, some institutions might not be ready for AI or may be working towards an institution-wide AI policy to establish a framework on how to leverage these tools. One other element of our AI Design Assistant is the ‘opt-in’ configuration. Institutions are always in control and must opt-in to utilize to the tool to ensure alignment with institutional policies and procedures.
Is there a role for AI in ensuring course consistency? Why is that important?
One of the most common themes in the feedback we hear from students is the need for consistency between their courses and across their higher education experience, but that inconsistency is all too common, making it equally confusing and frustrating.
Students reference issues like where they can find information, the navigability of a course, transparency about how the course or grading will work, or what the expectations are for an assignment. That can vary significantly from course to course and professor to professor. That is a question of course design and is not going to be solved by AI, but it can play a role in smoothing out some of those issues.
Developing grading rubrics is another example. Grading rubrics can help provide more consistent and more transparent grading and feedback. The AI Design Assistant in Blackboard Learn can help instructors create those rubrics more easily and quickly. It is about creating a consistent student experience, while still giving instructors freedom and flexibility.
How does authentic assessment fit into the overall conversation, and what can educators and institutions do to properly align testing to the challenges presented by generative AI?
When ChatGPT was first released, there was immediate concern that students would be using it to take exams, draft essays, complete assessments, or otherwise cheat academically. As a result, there was a lot of focus on AI detection. But after researching it extensively, we concluded that AI detection is just not accurate enough and poses significant student experience and inclusivity issues.
A more valuable approach is helping instructors create more authentic assessments. Authentic assessments prompt students to engage in critical thinking, draw upon firsthand experiences, and contextualize their knowledge within current events. While this approach fosters deeper learning, it also poses a greater challenge for AI utilization.
Once again, our Blackboard Learn AI tools prove invaluable in this regard. We’re continuously enhancing its capabilities to offer instructors recommendations for designing authentic assessment activities. It is ironic that we are utilizing AI to raise the bar for students’ use of AI. However, this approach is firmly rooted in sound pedagogy. Instead of solely focusing on detecting potential cheating on assessments, our aim is to cultivate superior assessment methods.
This is an exciting idea, because the field of higher education has been pursuing the goal of more authentic assessments for a long time, and with the help of this innovative technology, we may finally be seeing real progress.
To learn more about Anthology’s research on generative AI and their on-going product strategy for topics such as Authentic Assessment, go to AI, Academic Integrity, and Authentic Assessment: An Ethical Path Forward for Education | Anthology