Is it cheating?

Everyone likes to add value, which is why new labour-saving devices can feel like cheating - right up until they become so mainstream you don't know how you ever found time to do it the old way. That's where we are with AI adoption, worrying about whether we're still doing our jobs properly if we use it. But you're not giving up your vacuum cleaner and getting the carpet beater down from your Nan's loft, and we're certainly not surrendering Word and getting the typewriter back out or unclogging that bottle of Tipp-Ex*. Let's save time feeling guilty and proclaim with confidence: I wrote this with the help of AI.

Having a good assistant isn't cheating.

Hand in your calculators

In 1966, a group of maths teachers protested against the use of calculators in class. Sitting here now - with near-as-dammit unlimited access to online tools, Excel, a calculator app on my laptop, and another one on my phone just in case - worrying that kids will get too used to having a calculator to hand feels quaint.

I would worry more if a kid left primary school still unable to use one. Beats using TikTok if you’re choosing gateway tech for your children.

To be fair to the teachers they were, I’m certain, more worried about the students than their jobs and they were making a fair-ish point about waiting until kids were older and had learned basic arithmetic skills first. I get that. But by ‘older’ they meant ‘about to go to college’ so let’s not cut them too much slack.

Fear of tech, fear of change, and a troubling lack of foresight seem like the bigger issues.

The plagiarism problem

AI can stir a similar reaction today, and there’s good money to be made selling tech solutions that promise to detect whether a student wrote their essay in ChatGPT. But promoting software that claims to do that in the name of plagiarism only compounds the sense that using AI is wrong, and that’s the problem. Using AI to generate content isn’t inherently dishonest as long as you're honest about what you did and how you did it.

The calculator comparison holds: you still need to know how to do some basic maths but AI is going to be – in reality already is – everywhere and you need to learn how to use it well just to keep up. And if you use it well, it’s your best study buddy. Like the calculator, it will help you crack through some basic but effortful things quickly so you can learn more, learn faster, move on the bigger and more complicated – frankly, more interesting - problems that require your uniquely human capability for working with other people, for creativity and for problem-solving.

And if you find it’s saved you so much time you can afford to mess about at the back of class for a minute or two then you can use your AI to turn that email into a poem. And kids can still get their calculator to write rude words, through the magic of turning it upside down. If it helps keep you cheerful and motivated, and teaches you something about innovation, it's time well spent.

07734.

372215. 8008135. 707.

376616.

The other teeny tiny problem with plagiarism detectors is that detection software doesn’t really work. Or, at least, not in the way you think. Tales of Turnitin woe abound on Reddit, with original work sometimes scoring high on the plagiarism scale – sections of A Tale of Two Cities apparently scoring around 70% - while work drafted with help from ChatGPT can score as low as 6%. Worryingly, false positives seem to disproportionately impact ESL and neurodivergent students who are more likely to write in a formulaic style. Software buyers beware.

Don’t get lazy

If you’ve used Chat GPT a couple of times you’re probably already pretty good at detecting it, particularly when someone’s done a straight-up lazy lift and drop of vague, generic content. It’s the Constant Capitalisation Of Every Word. It’s the neat regularity of subheadings. It’s all those bullet points and even – shockingly often - the unchanged font and US English spelling coming from a British colleague. Then there’s overuse of certain words and phrases: “I hope this email finds you well… utilize this, deepen your understanding of that, in conclusion..." Mainly, it's because it's not quite right, not personalised and not relevant or specific. You don't need plagiarism software to point that out.

Cheaters gonna cheat, as Taylor Swift didn’t say. If you’re too lazy to adapt ChatGPT's first response to your no-context prompt, or you’re happy to hand over something super generic and then not inclined to be open about how you used AI, then you’ll get caught out sooner or later. People will stop replying to your emails, you won’t win the bids - and your essay will not pass muster.

At that point, you might as well copy off the kid at the next desk, without really knowing what you were writing down. And your teacher can spot that a mile off.

Make it your own

Let AI assist you with generating Your Work. Play with prompting and you'll quickly become your own prompt engineer.

1. Train the AI on your unique style, ensuring (another classic ChatGPT word) the output reflects your voice. Tell it what that is. Better yet, get an in-house AI that always writes in your voice.**
2. Answer the question you're answering with information that you found and verified and make sure it’s relevant to the question - in other words: read the AI content back and make sure it's right before you share it. A bit like checking the output of your calculator to make sure the answer it's given you isn't wrong by several orders of magnitude! Make the content less generic by incorporating your perspectives and experiences. If you have access to a secure AI**, feed it your materials to create contextually relevant and unique content only.
3. Go back and forth a bit and use the right tools for the right things. Sure, Copilot claims to do it all but in practice Perplexity is good for finding information and letting you verify your sources, ChatGPT produces quality drafting, Canva is good at graphics, Gamma makes a lovely slide deck, Pi is great for bouncing ideas around… Have all the tabs open. Better yet: buy a nice, affordable and totally bespoke RAG GPT hybrid that does just the job you need doing, and does it well.**

‍Make your bed

Admiral William H. McRaven’s viral TED Talk taught us that accomplishing small tasks helps you work toward your goals and build a sense of achievement and purpose. It’s why we prefer cake mixes when we have to add an egg. It’s why your ready meal feels more like proper dinner when you put it in the oven not the microwave. The motivational power of accomplishment is baked into everyone, and you don’t get it from lazy AI cut and paste jobs. You do get it, though, if you train and prompt the AI to do something unique to you and you're happy telling everyone how you did it. You get it when ChatGPT responds to a good prompt or your comments on its draft and makes something better – especially if you’ve told it something it didn’t already know. Something about you.

*I should explain Tipp-Ex for younger readers, although if you don't know what Tipp-Ex is you're probably already using Perplexity, so really this is in case the WiFi is playing up: Tipp-Ex is a tiny bottle of claggy white paint with a chalky finish that people used in the 80s and 90s to hide typing errors on paper really poorly so they could write or type over them. Or you could use it as nail varnish to pass the time in double physics. You can still buy it, but lord knows why you would.

‍

**Shameless plug but for very good reasons: Engine builds private AI tools that let you keep things simple and honest; private GPTs trained on content you curate and your brand guidelines, giving you answers with citations that your teams can verify. We charge a one-off, affordable build cost and a single, fixed monthly cost for your whole organisation. Email us at info@engine-ai.co.uk.

‍