ChatGPT is not a threat to authentic assessment

Artificial Intelligence is not the problem. It’s our education system that needs to become intelligent, and fit for purpose at last. ChatGPT could be the kick that we’ve been waiting for – an impetus that could propel us into a more engaging, useful, and valid form of education.

Most states in Australia have now banned the use of ChatGPT in schools, amid rising panic about the implications of Artificial Intelligence on assessment, writing skills, and critical thinking. This is largely because ChatGPT threatens the validity of our assessments and assignment work, giving students the opportunity to avoid doing their own research, writing, and sometimes even dodge sitting their own exams.

The truth is, though, that the threats exist because our assessments are already invalid. Our tasks for students are all too often designed not for maximum educational impact, but rather for ease of processing, ease of marking, and ease of authentication. When a student sits in front of you, writing on a piece of paper, you can be pretty sure the exam is their own work. But there’s very little evidence that exams are actually a worthwhile assessment of a student’s learning.

Let’s face it, when you choose a doctor, you don’t want one who’s excellent at sitting exams. You don’t care about their ATAR. You want a doctor who has great communication skills, is an excellent diagnostician, and works collaboratively with other doctors and allied health professions. None of these things are effectively tested in exams, or measured by an ATAR. Nor can they be replicated by ChatGPT.

If our assessment of students’ learning and skills is based on assessments that ask them to regurgitate easily tested facts and processes, we’re not testing anything of value.

This isn’t because teachers don’t want to do better – it’s because they’re not given the time or resources to innovate. With insanely high workloads and class sizes, teachers have little choice but to do roughly the same thing they did last year.

It’s past time to transition to a problem-based, authentic style of learning and assessment that builds students’ critical thinking, ethical decision making, and problem solving skills.

While writing Raising Heretics, I did a trawl of University websites in Australia and overseas, and found they all had very similar lists of “graduate attributes” – qualities they claim their graduates have. They usually emphasise problem solving, ethics, creativity, and critical thinking. Sometimes they mention leadership. Not exactly characteristics we typically test in exams, are they? Not easily tested under exam conditions, either. Not properly. And why would we want to? We rarely solve problems in the real world under exam conditions. If we want to know how this person is going to perform in the real world, exam results are all but useless.

So how can we change assessment to be useful, authentic, and relevant? Simply by making the work useful, authentic, and relevant. When students are solving real problems, we can engage with them along the way, so that we know their contributions, and can identify where they are highly skilled, and where they need more support. We connect their learning to the real world by empowering them to change that world by solving problems in their own communities. This way, students tackle challenges without textbook solutions, and develop skills that can’t be assessed with multiple choice questions.

They also learn that they are powerful, and that the skills they are working with are worth learning, because they make it possible to effect positive change in their own worlds. Instead of copying down notes from a powerpoint presentation, or checking the answers to their work in the back of the textbook, students now have to evaluate their own work, challenge and critique their own solutions, and make the case for why their results are valid.

Suddenly authenticity is built in. Critical thinking is essential, as is effective collaboration – real problems are not easily solved by lone wolves. They need diverse teams to ask difficult questions and challenge assumptions.

Some schools are challenging the ATAR, arguing that ranking students according to academic results is pointlessly one dimensional, and they are right. The idea that numeric results based largely on exams are objective and unbiased is flawed, and we’ve known that for years. Postcode and ATAR are far too correlated for the ATAR to be valid. We need to go further than ditching the ATAR, we need to change our whole curriculum, to move away from facts and known processes, and finally build an authentic emphasis on problem solving, critical thinking, and ethics. ChatGPT can never match that.

Leave a Reply