Maps, Memory, and Machines: My AI Assistant and Me

The Joys (and Leaks) of AI Assistance

12 October 2025

Fliss Falconer

The Joys (and Leaks) of AI Assistance

I joke that Microsoft should pay me commission for every Copilot recommendation I make but beneath the humour lies a deeper truth: AI has become my creative co-pilot, my logistical lifesaver, and occasionally, my soggy tent enabler.

It makes lists. It builds ideas. It takes first attempts and polishes them up. It takes final drafts and aligns them with marking criteria. It can plan a three-day camping weekend and create a list of what to take, how to cater, and where to visit. Shame it didn’t remind me to check the 25-year-old tent for leaks before we went on a weekend with torrential downpours. It overhauls. It suggests. It is amazing.

It also signs you up for things that you’ve forgotten about.

Map-reading before SatNav: My Analogy for AI

As part of my research into AI, my favourite analogy for using it is that you’ve got to be able to read a map before you reach for Google maps to direct you. For those of us who have experienced first-hand the introduction of Google planning our routes, we like to think that were it to falter and stop, we could still find our way. We have developed the synapses in our brain on where we need to go, on which roads we must travel and how to navigate from a closed road (last weekend’s fun!).

AI is similar - if we can do it ourselves longhand, but the use of AI will free up our free time, then I am all for it. I am wary of the pitfalls that happen when I do not read a document through before I put it through Copilot first. If I had not scanned the text, then I will not have retained enough of the original and may lose sections or be blindsided by areas I had forgotten had existed.

It is not the be all and end all.

But it is bloody marvellous.

When Copilot writes you up (and signs you up)

So, I have been merrily letting Copilot be my upbeat and enthusiastic assistant for more than six months. I know it’s been at least that long because I received an email to say that I would be delivering my ideas about awareness-raising for our upcoming show ‘Princess Ida’ that’s being performed at the Courtyard Theatre, Hereford from 26-28 March 2026 (Plug, plug, plug).

My ideas? What ideas?

As part of the company’s questionnaire in April, I had some ideas about what I could do to pitch in and help to raise awareness for our wonderful singing group. I started with my ideas! But me being me, I asked Copilot to come up with further suggestions. I liked them. I sent them all in. At no point did I conceal that Copilot had helped me. For one, when I opened the email this week, not only is it filled with emojis, which makes it a dead giveaway, but I’d even done a crap job of removing Copilot’s own conversation pieces at the top and bottom (facepalm).

Reading them through, I can recognise which were mine but there were some that I did not recognise. It wasn’t my work. It wasn’t in my memory bank on which to pull.

This is crucially why we need to make it really clear to people early in the education process what plagiarism is, and besides the fact it is ripping off someone else, it physically debilitates you with being able to be called on, questioned or develop the original material.

Plagiarism, Memory and the Ethics of AI

Case in point: I call up a Google search (I like to do things the old-fashioned way every so often!) and look for some research articles on this.

No sooner do I click on what looks like a reputable source, then an AI bot pops up offering to do the hard work for me, under the guise of ‘strategic’ use. It is. Having access to AI has helped me with my university course so far, as I have been able to ask Copilot to help me locate which chapters would best inform me about emotional intelligence and the Kolb theories, and that has saved time. Taking that option away, though, and I would have been able to find it myself. It would have taken me longer, but I could do it.

Again, look at these blogs - once, when playing with my new tablet, I was getting to grips with ChatGPT and it planned a blog with me. For me? Even for a first attempt, it read as though I had written it. It planned my beginning, middle and end. I felt dirty adapting one or two lines it had suggested. Like doing a back-alley deal!

Anyway, the source by Himendra Balalle and Sachini Pannilage (2025) suggests that it challenges the integrity of the institutions that allow its use up to the value of the qualifications they offer. It tests the ‘quality of education that represents honesty, trust, and ethical conduct’. Like the gun theory that guns don’t kill people, but people kill people, I irreverently make the link that AI is only going to be as ethical as the humans that use it.

In 2023, in a conversation with ChatGPT itself, Geoffrey Currie questioned how it could be misused. Its reply included: ‘Like any other technology, ChatGPT has the potential to threaten academic integrity if used inappropriately... students may use ChatGPT to generate essays, assignments, or other coursework without properly citing or acknowledging the source... students may use ChatGPT to find answers to exam questions... students may rely too heavily on ChatGPT to provide them with answers and information, rather than engaging in critical thinking and analysis themselves, hindering their ability to develop important skills such as problem solving and creativity.’

Currie shows that this is nothing new, and this is corroborated by Balalle and Pannilage, who give a potted history of some very famous plagiarisers, Shakespeare being one of them. ‘Roman poet, Martial, who lived from approximately 40 AD to 104 AD, identified that his poetry was being copied and recited by other poets without his knowledge.’ ‘Plagiarism’ as a term was yet to be defined but the indignation of work being stolen without proper citing is as much a part of our modern world as it ever was. There is a keen culture on Facebook where citing posts from other sources is made clear in hyperlinks and hashtags, and woe betide if you forget.

It is going to have to be monitored in line with any other type of plagiarism. As Currie says, it is an issue with all technology - that it can be misused. There is always a loophole to be found. It is part of our humanity - curiosity and survival. If the technology was taken away, students will always find another way. To that end, Turnitin (the tool that detects the use of plagiarism in all forms for academic assessment submissions) has come on in leaps and bounds since I last was getting qualified for something. Thankfully, I suppose, there is as much of a rise in cyber police as there is with cyber criminals. And therefore, as much bureaucracy and scale in all.

My promise to myself

As I am maintaining my own integrity, I will not be using AI to help me write my essays for my course. I want to feel that I have deserved the qualification when it is finally mine (in 2028!).

I will, however, continue to be scuppered in my personal time, when it comes up with amazing awareness-raising ideas on which I have two weeks to prepare and then present.

Copilot... now, when we talked about a flashmob in town, how can we develop this with a flyer and a ticket-connected giveaway...?

(P.S. I like ChatGPT but I still prefer Copilot; Gemini thinks I’m an idiot, so I tend to only ask her for favours if she is in the mood.)