“It’s like I’m smelling you, but you’re also holding a bag of nuts,” my husband told me with a completely straight face. It’s not exactly what you want to hear when you ask your partner whether they like your new perfume. But then again, I did ask.
It’s also not as if I was completely sold on the scent, which was simply called “Katie Collins 03.” It had been designed for me, along with Katie Collins 01 and 02, by an algorithm at the VivaTech show in Paris last month.
I’d filled in an online questionnaire, and then submitted it to the Netherlands-based art and technology company EveryHuman, which makes both the tech and the perfumes. I then watched as three little bottles traveled along an automated production line, stopping regularly under different canisters of liquid, where the different elements of each scent were drip fed into the bottles. Once they reached the end of the line, they were packaged into a cardboard container for me to take home.
We’re increasingly seeing different ways in which AI is creeping into all aspects of life, with one of the technology’s most appealing promises being its ability to offer us more-personalized products and experiences. At VivaTech, I saw how big companies such as L’Oreal and Sephora were channeling this into helping people make better purchasing decisions, but I was especially intrigued to try out EveryHuman’s robotic perfumery, which would provide me with a custom-made product designed by an algorithm.
The beauty industry is as old as time, but over the past 10 years, cosmetic companies have been blending beauty and technology to provide better purchasing experiences and better products, to satisfy the growing demand for more sophisticated, personalized and sustainable fare. At VivaTech I also saw an infrared hair dryer, bioprinted skin and an app that works like a stylist.
My algorithmic perfume.
When it comes to choosing a perfume, I’m in need of some help, so EveryHuman’s fresh approach appealed to me. I’ve long struggled with finding fragrances I really love, but I’ve recently fallen for the buttery vanilla scent called Bianco Latte from classic Italian perfumer Giardini di Toscana. Still, at around $150 per bottle, I can’t justify it being my everyday go-to perfume. So I was very much looking forward to outsourcing my next new perfumes to EveryHuman, which designs three scents for you based on a personality test. The experience costs around $25 and is currently available in London, Vienna, Dubai and the Dutch city of Breda.
The EveryHuman quiz begins by asking you questions about where you grew up and what you like to do for fun — classic first date chat, in other words. But it quickly descends into having you rate, on sliding scales, your different attributes (like how talkative or forgiving you are), which introduces significant ambiguity into the process and feels like you’re sliding fast into therapy territory.
Once you’ve completed the quiz, which is thorough and takes about 10 minutes, the machine gets to work bottling your scents. I was keen to try out the perfumes on the show floor, so I sprayed all three on myself, which to be honest only added to the sensory overload of being in a busy convention center packed with people. I smelled… weird. There’s no other way to put it.
And so I decided to bring them home and try them on one by one for my husband, to see if he could offer some objective feedback. Here’s how it went.
Katie Collins 01, he said, “reminds me of a Victorian-themed speakeasy with vintage tweed furniture where I would be able to order a smoky drink served under a bell jar — and I quite like that as a concept.”
No doubt he was picking up on the notes of pepper and leather that were included in the perfume, which ultimately made me feel like it was a little too masculine for my liking. “I wouldn’t say alluring, but I don’t dislike it,” he added.
Katie Collins 02 he immediately identified as “fruity,” which could well have been the notes of peach and raspberry amid the tonka, musk, amber and sandalwood. This was the perfume with the longest list of ingredients of the three, and it also elicited the longest answer.
“I could imagine being out, with both of us wearing beautiful clothes — you know, you in a lovely cocktail dress, me in a nice suit,” he said. “That is a smell to wear when you’re out wanting to be seen.” I figured this was more of a success, but it made more of an impression on him than it did on me.
I asked my husband whether any of the scents were alluring.
Katie Collins 03, my personal favorite, he described as “very almondy” and “legitimately nutty.” If it wasn’t obvious, this is also the perfume he was smelling when he described me as a bag of nuts. It was an interesting reaction, given that the main ingredients were actually vanilla and jasmine. He noted that he preferred scents 01 and 02, but I didn’t feel a strong connection to them in comparison to 03.
As you’ll know if you’ve ever experimented with AI in any context, sometimes it gets things right and other times it totally misses the mark. Overall, I’m not sure that the novelty of designing perfume by algorithm quite lives up to its promise — but that’s not to say it didn’t teach me anything.
It turns out the common denominator between the scent I already wear and the one I liked the most from my algorithmic experiment seems to be the inclusion of vanilla. Hopefully, when it comes to buying perfume in the future, I’ll have a stronger sense of what I really like and what really suits me, so that I don’t end up smelling like a Victorian speakeasy.
+ There are no comments
Add yours