AI and marking


It was one of those things that you never forget.

I was teaching Grade 9 Business and the students had time experiencing Data Processing, Accounting, Marketing and Office Practice. It was in my Accounting section where I gave this question …

“Why do you use pen when you do bookkeeping?

And one memorable response was …

“Because if you used a pencil, people could arasit”.

Sound it out.

Recently, Alfred Thompson, Peter Beens and I had a chat about this.

Peter offered an assignment that he used with his students related to the concept. It’s linked here.

As computer science teachers, we kind of nerded out on how to interpret this. In my mind, it’s kind of easy since, while it’s encrypted, exactly all of the letters are in each of the words. There are only so many combinations and you could run them against a dictionary.

What if there were words like “arasit” though?

I’ve been hanging on to this article for a while now thinking about the implications.

Could exams be marked by AI? Ofqual launches ‘exciting’ competition

As far back as I can remember, this has always been the promise of technology. It’s exciting to think that a computer could do the assessment/marking and leave more time to the teacher to do the actual teaching and class observation.

I remember a science department where all of the marking was done on bubble cards. The philosophy was that assessments should be completely objective. Of course, there’s the questions themselves. We know that there really is an art to creating questions.

Not all assessments can be distilled to this format though. And, they never should be. Students should be equally assessed with their abilities to communicate and create original responses and that doesn’t come down to choose a, b, c, or d.

That’s why I’m particularly intrigued about the concept in this article. What would it actually take to write a piece of software that could assess an essay or a piece of artwork or a computer program, for examples? They’re substantially more sophisticated.

From my perspective, it’s way in the future. But, I’m not brave enough to say never. Given what’s possible with Artificial Intelligence today, it’s not inconceivable that it could happen as the science matures.

Just not anytime soon. It will be interesting to follow this competition and see what turns out.

OTR Links 03/11/2020


Posted from Diigo. The rest of my favorite links are here.