A man jumps into a taxi and starts chatting to the driver about philosophy. The taxi driver turns to him and says, “Are you a Catholic or a Protestant?”

The man says, “Well, I'm an atheist.”

The taxi driver thinks for a moment and says “OK, but is it the Catholic God or the Protestant God that you don't believe in?”

I've written before about the UK's rapid loss of faith, but it is always interesting to have statistics confirmed by lived experience. A few weeks ago, I was filling in a form that wanted some diversity information from me. You know the drill, rough age, race, disabilities or lack thereof, and religion.

This was their drop-down menu for "religion":

I find it interesting that all the denominsations of the world's major religions get lumped into one grouping. It doesn't matter if you're Northern Conservative Baptist Great Lakes Region, or Northern Conservative Baptist Eastern Region - you're just "Christian".

But if you have no religion? Well! You've got a choice of three! You're either a non-believer but don't want to put a label on it, or a fully paid-up member of the Dawkins appreciation society, or a wishy-washy milquetoast who doesn't want to offend anyone.

I was a little miffed that I wasn't able to register myself as an "Apathist", but I got over myself.

I've previously written about the social construct of race. The same is broadly true of religion. In some communities, the exact sect you're in matters. In other communities, you're just "other".

Can you imagine what that drop-down menu will look like in a dozen years? It's easy if you try...

Never Tell The Candidates The Rules Of Success

I read an infuriating blog post recently: "If someone doesn't send a thank you email, don't hire them".

This is a great way to limit your talent pool. Personally, I've been hiring for a dozen years, and I've never received a thank-you email. So we can already show there is a cultural disparity here.

Did you tell the candidates that they needed to send a sycophantic email to succeed? If not, why not? Is it any different to telling them what the dress-code is? Or that they need to prepare a presentation for the interview?

Perhaps there are some things that you think everyone should just intuit? What biases and assumptions are you making about the people you want to hire?

If you have hidden rules for getting a job, you're deliberately ignoring candidates who don't share the same cultural background as you.

Make Up Justifications For Not Hiring People

These are all things I've heard interviewers say about their "unique" and "quirky" interview techniques.

• "Whenever I interview someone, I always mispronounce their name. I like to see how they react to that."
• "When interviewing, I go for lunch with the candidate. I ask the waiter to subtly get their order wrong. A great test of the interviewee's character!"
• "We always leave candidates waiting half an hour before calling them in. How do they cope with delays?"

OMG! No! Let's ignore the rudeness (and potential poisoning of someone on company time). Here's the problem with these "tests".

You haven't set the success criteria. Do you want the candidate to be deferential? To be confrontational? Do you want women to laugh it off, but men to silently seethe? What's the correct response.

Most often, the interviewer just wants a power trip. And because they didn't set the rules of the game in advance, they get to make arbitrary decisions. They hire the person they like best and then make up a post-hoc justification based on the candidates' reactions.

Side bar - names

If you mispronounce a name, you're immediately telling the candidate that people with their cultural background are rarely hired, or spoken about, in this company.

Why would people want to work in such a homogeneous culture? Why would they want to work for someone who couldn't even be bothered to search the web for "How to pronounce Nguyen"?

Be Inconsistent In Your Questions

A few years ago I went for a job I thought I was well qualified for. I didn't get it, so I politely asked for feedback.

"The whole team thought you were great. You were probably the strongest candidate overall, but we were looking for someone with experience in XYZ."

I paused for a moment. They continued. "So we hired someone with 6 months experience, because we needed to hit the ground running."

"Actually," I countered, "I have two years experience in XYZ and have taught several training courses on it."

"Oh!" They said shocked, "You never mentioned that."

"That's because you didn't ask me any questions about it!"

Now, maybe that's my fault for not guessing their secret desires (it isn't). Or perhaps I should have read the job spec more closely (it wasn't on there). Or maybe they could have asked the same questions to all candidates (they didn't).

It's tempting to treat job interviews as casual conversations and have a meandering and natural chat.

Don't do that. A job interview is a structured process designed to let you consistently evaluate multiple candidates. If you are asking each candidate different questions, that's not a fair test.

Bonus Rules

A job interview isn't a set of trick questions. You're working collaboratively to see if you can work in the future - not trying to prove your intellectual dominance. It shouldn't be a test to see if they've read the same interviewing books as you.

You aren't doing the candidate a favour by interviewing them. If you select for deference, don't be surprised if you only get obsequious underlings.

There's no point hiring on "cultural fit". If your culture isn't strong enough to handle a little change, or challenge, then it simply isn't sustainable.

What other people think

It's not just me who finds these hidden rules annoying!

Siân Thomas

@drsiant

Replying to @AngharadStone@AngharadStone @edent Ha - me too Angharad - generally answer to anything - including my alter ego Stan!

---

Tim Hobbs

@timhobbs

Replying to @timhobbs@edent I hadn’t thought about it until now but I realise that when hiring we approach it with a mindset of “trying to find a great person to help us”, not “offering someone an opportunity”.

---

James Wood

@laMudri

@ianwalkeruk @edent The good manners to not clog up people's inboxes with useless emails, perhaps

---

Do you have any interview gotchas that you hate? Let me know in the comments!

It was a bizarre job - one I've discussed before - but today, wouldn't we just throw an AI at it?

I recently read "How AI/ML algorithms see nudity in images - Comparison of image moderation APIs by Microsoft, Google, Amazon and, Clarifai" - a blog post by DataTurks.
In it, they use a training set of images - a mix of "safe" and "nude" - to see how accurate various automated image moderators are.

I downloaded the dataset (I did this on my personal computer - not my work laptop!) It is a dreadful dataset and is an amazing example of how poor data can give misleading results.

Lack of data

There are 180 images in the set. 90 nude, 90 innocent. That's not enough images to do a proper comparison. Even if only one nude image is misclassified, that will dramatically change the reported accuracy of the system.

For a test like this, you need to compare hundreds of thousands of images otherwise small errors give a distorted impression of reliability.

Lack of diversity

If you look at the provided sample of "safe" images, you may notice something odd. All the women have blonde hair and fair skin.

I'm not going to get into a discussion on ethnicity and the Fitzpatrick Scale - but out of the 90 nude images, I counted only 15 images featuring non-white women.

By comparison, there were around 31 blondes and 7 redheads. Does that sound representative of the world?

Lack of diversity

To quote Kryten:

The nude data set only features women. Male nudity is, apparently, not worth considering.

Perhaps that's why so many of my friends get unsolicited dick-pics? No one is training the data correctly?

Copyright and Consent

Who owns the copyright in the images? There's no provenance provided. While some of the photos are clearly professionally shot - others look like they are stolen selfies.

Do people know that their intimate photos are being used to train and evaluate machine learning models? Can they remove their consent from being part of this data set?

Source

This database of images is called Yet Another Computer Vision Index To Datasets (YACVID) - and was last updated in 2009.

I hope no one is using it for training an AI. But I'd also argue that it is unsuitable for evaluating AI performance as well.

• Thanks for inviting me. Can you let me know if the venue is wheelchair accessible?
• So excited to be there. What are the crèche facilities going to be?
• Looking forward to speaking. Will you be providing a sign-language interpreter?
• Can't wait to see you. Please can you ensure that videos of the sessions have subtitles?
• Here are the slides you requested. Does the venue have gender neutral toilets?
• Cheers, I've booked my travel. What's the link to the Code of Conduct?

These aren't things which I necessarily need - but they're incredibly important for people who might feel nervous about making special requests.

I'm not saying you have to do all of these - or that these are the only issues you should care about. But pick something. Spend a few of your privilege points on making the event a bit more inclusive for everyone who attends.

Huge thanks to Charlotte Jee and Jemima Gibbons for encouraging me to write this post.