The ultimate guide on using AI to ace an interview at Monzo

[mehmetyilmaz]

Hello! I’m Mehmet and I’m a recruiter in the COps Hiring Team

As we all know, AI-usage has surged everywhere and within hiring we are seeing tons of candidates use AI to aid them during different parts of the hiring process.

We’re super open and adoptive to candidates using AI - but we want them to use AI tools whilst still being authentic and being their true self

For example we do unfortunately see lots of examples of candidates simply generating answers to questions and pasting them directly from tools like ChatGPT. This means their application doesn’t stand out and lacks that human approach which we love and want to see

We’ve written guidance on the best ways we think you can use AI in our hiring process.

The ultimate guide on using AI to ace an interview at Monzo

By outlining this stance, we hope this will guide candidates and also show us as being adoptive and forward thinking. There’s been case studies of other companies who have banned AI use or had a hardline approach towards AI being used which has lead to bad press, and also isn’t very inclusive.

Happy to help with any questions on this and open to everyone’s thoughts as well

[JIMMWX (Village Idiot)]

Thanks for writing this. I see this increasingly more when looking to grow our directorate, it’s really frustrating when a candidate has clearly used A.I

[mehmetyilmaz]

Thank you! This is my first time posting in the community so appreciate the help

[mehmetyilmaz]

@ndrw

Is there anything being done to unpick people with a genuine interest (but perhaps more limited application) vs people who talk about everything Monzo do but got their info from AI?

It’s a great call out - we want to make sure its clear firstly that using AI is completely optional and we’re not assessing people based on whether they’ve used AI or not. From experience, it’s very clear who has used AI and who hasn’t, and in fact those who haven’t used AI are standing out more right now as their application answers sound more authentic, which is what we recruiters love to read. If someone has used AI to gather industry research for example or more info on what we do, but they add their own personal touches, that’s something we’re comfortable with, it shows they’ve done some research but have taken the care to add their own input. Also, the onus is on us to ask questions where answers can’t be easily generated by ai-tools and require candidates to talk on their own experiences. We can’t stop candidates from using AI of course and we don’t want to, but its definitely our responsibility to separate those who have put in genuine effort and those who have just copied a bunch of info - which I think we’re quite savvy to already - if someone has blagged it then they usually will struggle in the interview later on anyway We have ai-detection tools too which aren’t always 100% accurate but can help us spot 100% ai-generated responses

[projectfortytwo (Chris)]

I’m so glad I’ll be retiring soon. I’m with Douglas Adams on this.

1. Anything that is in the world when you’re born is normal and ordinary and is just a natural part of the way the world works.

2. Anything that’s invented between when you’re fifteen and thirty-five is new and exciting and revolutionary and you can probably get a career in it.

3. Anything invented after you’re thirty-five is against the natural order of things.

[mehmetyilmaz]

@N26throwaway

I’m wondering if there’s been a recent shift in Monzo’s perspective. Is this a new stance, or was this always your stance?

When it first came to light that Monzo were using AI detection with job applications, it was accompanied by the connotation that you weren’t at all accommodating of folks who used it. Simply discarding applicants when you detected any use of AI.

Yes when AI first ‘burst onto the scene’ about a year ago, some of us were definitely more skeptical about candidates using it, and some of our adverts asked candidates not to use it, this is correct. We’ve learned a lot more about AI in that time and needed to educate ourselves within the team on the benefits of its usage. Now one of the reasons why we’ve introduced the blog is to align the whole hiring team on our stance now moving forward

[mehmetyilmaz]

@N26throwaway

How do you know with any reasonable degree of certainty?

The detection tools available just don’t work reliably, and more often than not produces false positives. And the generative models are evolving far faster than detection algorithms can keep up. I don’t think they’ll ever work reliably.

In academic contexts, we can’t do this reliably with coursework yet in all but the most extreme obvious cases. Academic courseworks also ostensibly train humans to write like machines. Or perhaps it’s because models were ostensibly trained on such writing styles?

In either case, for many people, their natural formal writing style closely resembles the tone of a ChatGPT output. I know mine does.

Yep we can’t say for 100% definite if someone has copied and pasted an answer, as the ai-detection tools aren’t 100% accurate you’re right - so we don’t just accept or reject an application based solely on whether someone has used AI. We do also guide candidates to our Writing Principles / Tone of Voice and ask candidates to consider this when writing their answers too

The statement that candidates are copying and pasting answers is a fairly confident assumption as I screen 100s of applications a week see trends of extremely similar answers, formatting of answers being identical and certain language or words used multiple times, so I start seeing trends which leads me to believing this

[hldn90 (Harry)]

I had my first AI interview encounter a few weeks ago. It was a first round over Teams. The candidate was clearly typing the questions into ChatGPT and then reading verbatim from her screen. I had a lot of second hand embarrassment. Didn’t really know whether I should have said something (obviously I couldn’t know for certain), but in the end just pretended I hadn’t noticed and wrapped it up fair quickly.

[BritishLibrary]

My other half was recruiting recently and had what she thought was a ChatGPT answer to part of the interview.

The part with a technical challenge element;

She was suspicious that the answer seemed a little canned, so ran her challenge through ChatGPT and was maybe 90-95% similarity to the answer/structure the candidate gave.

Super interesting to start to see this though - both in a ahhhh dystopia way as well how “how do I use AI to polish up my interview prep a bit”

[mehmetyilmaz]

Yep we’ve encountered this too unfortunately and have made a point to discourage it in the blog interviews are all about building a connection I personally on the call would try advise the candidate not to read directly off a screen and to keep it conversational, but no doubt it’s definitely a tricky one to handle