Artificial intelligence is creating new opportunities across higher education, but it is also forcing institutions to make difficult decisions. How can they embrace innovation while protecting students, supporting educators and maintaining trust?
For HOGENT (Hogeschool Gent / Ghent University of Applied Sciences and Arts), a university of applied sciences in Belgium, that question became impossible to ignore. After making the switch to D2L Brightspace, they were also grappling with the growing role of AI in education. Rather than approaching those challenges separately, HOGENT saw an opportunity to rethink how technology, teaching and institutional values could work together.
Their conclusion was simple: successful AI adoption starts with policy, not technology.
Starting With Principles, Not Tools
Like many institutions, HOGENT found itself facing increasing demands from a growing digital ecosystem. Maintaining its existing learning environment meant keeping pace with accessibility requirements, cybersecurity concerns, mobile experiences, system reliability and a rapidly evolving technology landscape. For a small team, it had become increasingly difficult to manage everything internally.
At the same time, AI was emerging as a transformative force across education.
“We were dealing with two transformations at once,” explained Stefaan Vanbillemont, Education Technology Specialist with HOGENT. “Changing our LMS and understanding how AI would impact education.”
Before exploring specific AI tools (like D2L Lumi) or use cases, HOGENT focused on the principles that would guide its decisions. The institution built its approach around respect for laws and regulations, responsible use, ethical and sustainable practices, and transparency.
Those principles apply not only to students, but also to educators and the institution itself.
“We ask students to be transparent about how they use AI,” Vanbillemont noted. “We should expect the same from ourselves.”
That commitment has helped create a consistent framework for evaluating new technologies and determining where AI can provide meaningful value.
Not All AI Uses Should Be Evaluated the Same Way
One of HOGENT’s key insights is that AI is not a single category of technology. Different uses come with different benefits, risks and expectations.
For example, an AI system that provides information must be evaluated for accuracy, reliability, privacy and accessibility. A tool designed to support instructors requires a different lens, one focused on educational quality, transparency and maintaining instructor control over course materials. Student-facing AI raises additional questions about pedagogical impact, equal access and creating safe learning environments.
This distinction has helped HOGENT move beyond broad discussions about AI and toward more practical conversations about where and how specific tools should be used.
Rather than asking whether AI belongs in education, the institution focuses on understanding the role a particular AI application should play and the safeguards needed to support it.
Using AI to Scale Support While Keeping Humans Accountable
Those principles have shaped HOGENT’s use of D2L Lumi.
One area where the institution has already seen significant value is support. AI-powered assistants help answer routine questions from students and staff around the clock, making it easier for users to find information when they need it while reducing pressure on support teams.
The impact has been substantial. Between September 1 and October 15, 2026, HOGENT’s three AI-powered support assistants handled 9,917 conversations, responding to more than 18,000 user messages with a 99% resolution rate. During that same period, only 103 conversations required escalation to a staff member.
For Stefaan Vanbillemont, those numbers demonstrate the practical role AI can play in higher education.
“If every support request takes five minutes to read, route and answer, those conversations would have required more than 826 hours of staff time,” he explained. “With the AI assistants handling the vast majority of requests independently, only about nine hours of human support time were needed for escalations.”
That difference amounts to more than 800 hours of potential staff capacity over just six weeks, allowing support teams to focus on the complex questions and situations where human expertise, judgment and empathy matter most.
The goal, however, is not to replace people.
Instead, AI acts as a first point of contact, helping users find answers quickly while ensuring staff can dedicate more time to higher-value interactions that require context, nuance, and personal attention.
This balance between efficiency and accountability appears throughout HOGENT’s AI strategy.
Keeping Educators in Control
That same philosophy guided HOGENT’s approach to course migration during its LMS transition.
Rather than automatically transferring everything from the old environment into the new one, the institution chose a more deliberate process. Educators are able to review existing materials and decide what to move forward. D2L Lumi can assist by helping organize content, generate activities or streamline development tasks, but instructors remain responsible for what ultimately appears in their courses.
For HOGENT, maintaining that human oversight is essential. The technology can make suggestions and reduce administrative effort, but the educator remains responsible for evaluating whether those suggestions align with learning objectives and student needs. Teacher control, transparency and professional judgment are treated as foundational requirements rather than optional safeguards.
Exploring New Possibilities for Learner Support
HOGENT is also exploring how AI can support students more directly.
Tools such as D2L Lumi Tutor create opportunities for learners to ask questions and receive support outside traditional classroom hours. Because these experiences are connected closely to course materials, they have the potential to provide guidance that is more relevant and contextual than generic AI tools.
At the same time, HOGENT is approaching these capabilities thoughtfully. Questions remain about how much students and instructors will trust AI-generated responses, what role AI should play in learning, and how institutions can ensure these experiences support rather than replace meaningful engagement with course content.
Instead of rushing toward widespread adoption, HOGENT is carefully evaluating where these tools can have the greatest positive impact.
Moving From AI Policy to Everyday Practice
Many AI conversations focus on the newest tools, features, and capabilities. HOGENT offers a different perspective: technology matters, but technology alone is not enough.
By establishing clear principles first, evaluating different AI applications according to their intended purpose and ensuring humans remain accountable for outcomes, the institution has created a framework that connects policy with practice. The result is an approach that not only guides decision-making but also delivers measurable impact, from freeing hundreds of hours of support capacity to creating new opportunities for learner and instructor support.
As higher education continues to navigate rapid technological change, HOGENT’s experience serves as a reminder that responsible AI adoption requires thoughtful decisions that support educators, empower learners, and reflect the values institutions want to uphold.
Written by:
