string(9) "edinburgh"

Flatmates not pulling their weight? I asked ChatGPT how to text them to let them know

Did it go to plan? Not really no


If there are two things that are inevitable whilst you’re at Edi, it’s that you’ll have a mice problem and that at some point you will come to blows with your flatmates as the share of the flat tidying is not being shared equally.

Nobody likes confrontation and there’s nothing worse than being stuck in a flat with flatmates to who you’ve just sent a passive-aggressive text accompanied by a picture of a full sink of dirty dishes… So with the power of AI now at our fingertips I decided to use ChatGPT to solve the student-wide dilemma of how to bring up tidying with your flatmates without potentially causing a nuclear friend fallout.

I prompted ChatGPT by initially asking it to “compose a text message to a university flatmate about helping out with tidying more in a passive aggressive way”. However, the benevolent AI then presented me with a preachy message about how approaching situations with aggressiveness is not the way forward and it would not help promote such behaviour. This should have been my first warning sign, AI wouldn’t last five minutes at uni.

Undeterred I repeated my request but removed any mention of passive aggressiveness and was greeted with another (slightly) preachy message about my request. The AI told me that “asking your university flatmate to help with cleaning the flat more often via text message can be a bit tricky, but it’s definitely doable”. It also told me to “be direct and clear about your concerns and what you would like to see happen. It’s important to avoid blame or accusatory language, and instead focus on finding a solution to the problem”. Who knew I’d get a free therapy session with the AI as well?

It gave me a prompt that I should use to send to my flatmates, but I also wanted to see what it expected my flatmates to respond with. I amended the request a final time, asking it to present the conversation with multiple flatmates and their responses.

Lo and behold it eerily created a thread of text messages which I would use to compare with my beloved flatmates responses. Praying that my flatmates would forgive me for my boldness I started typing word for word what the AI told me to. At this point, my living situation was quite literally in the hands of a robot.

Lo and behold it eerily created a thread of text messages which I would use to compare with my beloved flatmates responses. Praying that my flatmates would forgive me for my boldness I started typing word for word what the AI told me to. At this point, my living situation was quite literally in the hands of a robot.

I pressed send and prayed. I consulted with the AI who predicted that my flatmate would respond with the following: “Oh yeah. I’ve noticed that too. I’ve been swamped with assignments and projects, and I haven’t had much time to clean up.” Honestly, pretty accurate for a standard uni student response.

However, what the AI hadn’t predicted is that my flatmates are angels who instead of making excuses immediately offered to make a cleaning rota, even going as far as to say they thought it was a “really good idea”.

Whilst the people pleaser in me wanted to respond in a nice way, I knew I had to continue with what the AI told me to. The issue here was that the AI assumed my flatmates would need more convincing so I ended up sounding like a complete d*ck.

Very understandably my flatmates felt attacked, and ChatGPT failed me as it had expected my flatmates to say that “you’re right, I’m sorry, I’ll try and make more of an effort”.

The icing on the cake was the final expectation from ChatGPT where at the end of the texts myself and my flatmates would “smile at each other”. Whilst I couldn’t see my flatmates reaction something tells me they weren’t smiling.

I then revealed to them these truly weren’t my words and instead those of an AI. They weren’t that surprised, I’m usually far too afraid of conflict to ever be bold in the group chat.

So what have we taken away from this? Well firstly, the AI definitely assumed my flatmates would go on the defensive and their willingness to immediately create a rota definitely threw it off course. But ultimately it’s all generic and an AI would never be able to guess exactly what your flatmates would respond with.

So, would I recommend using ChatGPT to help with broaching an awkward situation? Honestly, yes! I tried some different prompts and if you’re not completely sure how to correctly phrase something or even politely tell your lecturer you’re not able to make the lecture it’s eerily good at creating a good template. Let’s just hope it doesn’t get too good or else it’ll start writing all our articles for us, or maybe it already is…

If you want to use ChatGPT to tell your date you’re not that into them, or send an eloquent email to your lecturer about your essay being late when you’re steaming at 2am then simply click here!

Related stories recommended by this writer:

• Edinburgh University staff to strike for 18 days between February and March

• Over a fifth of Edinburgh students say they are addicted to Elf Bars