While doomscrolling earlier this week – bad habits, but hey! – I caught the Aldi supermarket thread (great doom scrolling!), where the question was posed (paraphrased): if other supermarkets were a sport what they would be? … stay with me. Amused by the comments, my scrolling continued. Apart from being distracted and mildly amused, I was struck by AI in the comments and how much the ‘happenings’ just totally resembled assessment issues and debated in universities right now. So, what can we learn? 

1. We seem comfortable with AI generated responses when they’re clearly instructed by human creativity. An AI generated image capturing what appears to be human good idea had no negative responses. Just good bantz. We seem comfortable with AI helping to realise human created ideas. Afterall it’s a tool to assist, right? Likes, laughs and affirmation result. 
2. We scorn unabashed unmodified AI slop. More than that – slop appears to generate negative comments on the environment (was this answer worth the environmental cost – it doesn’t add value?), observations that AI lacks humour, and AI isn’t adding to the thread. I get this … if someone presents me with an unmodified AI response, I immediately feel negative. My baseline tolerance of the downsides is already alert. 
3. The lazy edit – content that looks and feels AI but isn’t quite provable – is tolerated but largely ignored. Posts that read as AI-ish – emoji-heavy, bullet-pointy, low on nuance, listy outputs – tend to get scrolled past.

So what might this tell us – AI gets a better response when is enhances a vision or human creative force; we all hate slop and raw outputs; AI like outputs lack impact. Moral of the story, develop your vision and use AI to reach it and develop it. Bring your humour, thoughts and ideas to the table and let AI enable you. Perhaps a critical doom scroll is an exercise for students when familiarising with AI – asking what is your tolerance? Scan the comments and see what feels OK.

Perhaps these threads say something about public exhibition too. In my unscientific doomscroll sample, it was clear that AI generated comments receive a more negative outcome (number of hits and sentiment) compared to the obviously human. Translate this to assessment … by making our assessments more public, perhaps the stakes are higher and the test of interaction, reaction, and impact matters more than the judgments of a single marker sitting alone in an endless game of cat and mouse. I have moved my thinking on authentic assessment over years, but I am increasingly convinced that the ‘audience’ aspect is important in an AI world for accountability, relevance, and the full realisation of creativity. I am not suggesting non-specialist crowd source judgments, but there is something about holding our work to a more public standard that could just help us be more aware of what we put out. 

AI reflection: No AI was used to write this – all errors are mine; but in a show of imposter syndrome, I did just check whether it made sense (ha ha). Please note – I use a lot more dashes than I used to … AI is changing how I write even when I don’t use it!