Stop, Improve, Start is a six-part series that looks at current trends and research to help faculty, instructors, and education and training providers meet the evolving needs of students. Three Things About AI in Education is the first instalment in the series.
- Stop avoiding serious conversations about using AI in your college or university. Artificial intelligence is making a real difference across several sectors — financial services, health care, manufacturing, retail — but it has yet to have a significant impact on education. There are many reasons for this, but we should be actively engaged in a purposeful, realistic conversation about what AI could do to help more students access learning and be successful.
- Stop deploying AI predictive analytics without engaging students, teachers and administrators in an active, focused discussion about ethics, privacy and security. There are real and significant issues in such a deployment, and transparency is key.
- Stop thinking that AI will replace teachers and lower the cost of education. It will not. First, AI chatbots (even with a neat animated deep-fake avatar) are no replacement for the relationships that teachers have with students. Education is a relationship business in which compassion, care and challenge are as important as cognition. Second, AI has significant initial upfront costs and takes time to deploy effectively. There are no quick fixes in education.
- Improve the way algorithms work so that we know not just what they are suggesting, but why. If the “system says” I am a student in danger of failing, then we should at least understand the basis of that prediction.
- Improve our collective thinking about adaptive learning. Move beyond seeing it as a remediation process. Different learners take different routes to the same learning outcomes and, in doing so, can shed light on what works and doesn’t work in our learning designs. Rather than see adaptive learning in terms of test scores leading to remediation, we should reimagine this as different paths to success. Few adaptive learning engines are getting this right — and there are some real failures. This can improve through collaboration.
- Improve assessment engines that leverage AI. The pandemic has led to significant changes in how instructors assess student learning, including more authentic assessment, more peer assessment and more team assessment. AI could enable even more rapid deployment of these approaches but seems stuck on a “test and trace” set of assumptions about what assessment is. It’s definitely time for a renaissance in AI-enabled assessment.
- Start deep and effective collaboration between EdTech AI developers, teachers and students to deploy AI effectively. Learn from the work of tech hubs, such as the EDUCATE EdTech hub in London (UK) or MaRS in Toronto and take the time to understand the real experience of engaged learning and authentic assessment.
- Start looking at specific collaborative developments such as Open Education AI Resources (OEAIR), especially those that are related to areas of skill shortages and trades. See AI as a community resource, not specific to a given institution.
- Start investing in AI that makes a difference to equity and inclusion, especially for Indigenous learners. Be creative in capturing the role of elders, the power of learning circles and Indigenous ways of knowing in the deployment of AI as a shared service.