Our latest webinar series, Accommodations in the Digital Age, came to its conclusion with part three of the series exploring the role of AI in education.
The session included discussions on the proliferation of AI, how much help is too much help and an exploration into how we can effectively collaborate with AI across education.
This is a topic that has recently come to the forefront of our collective conscience, most notably with the rise of ChatGPT. Yet, AI has been with us since the 1990’s through technologies such as spell checkers.
However, recent studies have showcased an inherent race bias amongst other challenges raising questions about the models on which these new and emerging technologies are built.
Such realities prompt debate around how much reliance on AI is appropriate, especially in education.
Here are three of the key takeaways from the session…
The undoubted proliferation of AI
72% of students said AI-powered study tools helped them feel more confident with difficult academic material (Carnegie Mellon University, 2021). A number which, if anything, has only increased over the past couple of years.
The availability of this generative AI means everyone can, in effect, create a personal tutor. Research shows tutoring is a high impact, yet often inaccessible education strategy; back in 1984 it was coined the 2 sigma problem. So now the obstacle of accessibility for all has been overcome, why wouldn’t you put this in the hands of students using an AI chatbot?
There are also a wealth of personalised tutor tools that will help in all kinds of student situations with Khan Academy being a prime example. They built technology that will not only test students and provide material related to their interests, but when asked why students should learn this, respond as a tutor, not just as an answer machine.
The underlying issue at the heart of such proliferation however is that of ensuring equitable access.
But even if this remains unanswered, there can be no doubt that AI underpins much of what we use today and, as such, we need to understand the different ways this technology is implemented to guarantee that the accommodation goal is to support and not shortcut learning.
How much help is too much help?
As much as AI has become part and parcel of the education landscape, educators are naturally asking how much help is too much help?
Ultimately, the goal is to understand how we can use AI technology to support students' learning and not just do it for them.
A useful comparison can be made by examining existing accommodations and their AI counterparts. For example, a student could have independent accommodations for peer note taking, assignment preparation and text to speech software, all of which could be achieved through the use of a single AI chatbot.
There is an argument to be made that making AI tools and functionality available to students increases inclusivity.
However, we need to make sure that we aren’t removing the parts of the process which assists with information retention and the learning process, but rather remove the wastes i.e., the bottlenecks which consume time and take up valuable cognitive load.
This underpins the two key principles we are currently working to around AI at Glean
It SHOULD NOT do things for the learner but SHOULD remove their constraints. And it SHOULD NOT shortcut the learning process but SHOULD make it more efficient.
Successfully collaborating with AI
So how can educators successfully integrate AI and our student cohorts without compromising their learning experience and abilities?
The answer lies in balance.
If we are giving access to tools that have AI capabilities, we need to be mindful of how these tools work, so we can educate students on their limitations.
More broadly departments or institutions need to be thinking about our policies and ethical position around the use of AI, so that when we are evaluating products we don’t just reject on principle anything which may have an AI label.
Institutions have long been developing student creativity, critical thinking, communication and collaboration skills. Collaborating and communicating with AI gives space to practise these skills.
But what about critical thinking? If AI is telling students information is NOT correct, are our students skilled enough to critically evaluate what has been shared?
In truth, we don’t know the definitive answer on what is good learning with AI as yet.
We can draw on what we know about learning as a whole, which is that summarising learned materials into your own words has a significant boost on learning and test performance.
Success isn’t measured in terms of the number or articles a student has read, rather how they have applied what they have read to their prior knowledge and what arguments they have formed. We look at how they have synthesised the information.
The human level of involvement in learning is still high, it is a personal process, and this is one of the huge advantages of AI, its ability to adapt to the learner and create a personal experience.
It is precisely because this technology can be used for both good and ill that education needs to be proactive about how it engages with AI and start to think about procurement and ethics standards and work with companies that align here.
With thoughtful standards and responsible implementation which keeps the learner experience central to all developments, AI presents a bright future for accelerating and enriching learning.
To learn more about the topic, click the button below to rewatch the session in full.
