Just as the United Kingdom’s lockdown was beginning, The Interline sat down – virtually at least – with Matthew Drinkwater, head of the Fashion Innovation Agency at London College of Fashion.

By any definition of the term, Matthew is a bit of a renaissance man.  As well as being a leading light fixed to the prow of fashion technology – itself a crossover discipline – he’s also collaborated extensively with LucasFilm’s ILMxLAB on real-time digital augmentation technology, and he’s sought after on the speaker circuit.  Although he’ll be doing most of his presenting digitally for a little while, which we imagine he’s well prepared for.

We set out to speak to Matthew prompted by the news that the Innovation Agency was now running a course that would help prepare LCF students for a future where technologies under the AI umbrella – such as computer vision – are interwoven with fashion.

As you might expect, we got side-tracked.

The Interline: Explain to us what the Fashion Innovation Agency’s objective is, and how you go about accomplishing it.

Matthew Drinkwater: We have a really broad remit from the college to go and explore any emerging technology and analyse its likely impact on the fashion industry.  In practice, what we like to do is take new technologies and build proof of concepts that can demonstrate their potential to the fashion and retail industry.

Those technologies tend to fall into three main buckets.  The first contains things that might change the way designers actually create their products and collections: smart materials, smart processes and so on.  The second is for technologies that can be used to showcase collections, and as you can imagine 3D has been a long-standing chunk of work.  For those technologies, London Fashion Week has been a very successful testing ground.  We’ve been able to deploy immersive technologies like augmented reality and mixed reality there.  And the final group is anything related to the retailing process.

So what we’re doing it taking a technology – whether it’s AI, blockchain, or 3D printing – and actually building something with it, to get a sense of what works and what doesn’t.

The Interline: So in a way you’re road-testing the next generation of technology before it’s ready for the open market?

Matthew Drinkwater: Yes.  We work with technology companies to pull out what’s currently in research and development, and still probably about two to three years away from commercialisation.  So what we’re working with, technologically-speaking, is right on the cutting edge.

The Interline: And where do brands and retailers come into the picture? Because it’s probably fair to say that if a technology is years away from the open market, it’s probably not on the average fashion company’s radar.

Matthew Drinkwater: With our proof of concepts, we always try to bring in a fashion partner.  That could be a luxury brand, an emerging designer, or in some cases a student designer.  Because, as you touched on there, as much as fashion wants to portray itself as an early adopter of new things, there’s a lot on the horizon that the industry just doesn’t understand – and that creates resistance.  It’s difficult for a brand to believe in something until they see it, and it’s hard to see it when it’s not yet being actively sold.

Take 3D as an example.  The integration of 3D has become more of a talking point in the last eighteen months, but even now there’s still resistance to doing it – and this is something that was ready to be embedded in fashion a long time ago.  Now that people can see 3D in action, in design, development, augmented reality and a host of other applications, they understand where its value comes from.

So our goal is to demonstrate to the fashion retail industry how upcoming technologies work in practice and what the real benefits are likely to be.

The Interline: You’re also on-campus, of course, so presumably that kind of experimentation runs both ways – into industry, and towards the next generation?

Matthew Drinkwater: Absolutely.  Because we work with some really exciting external partners, we’re able to get hold of up-and-coming stuff – whether it’s hardware or software – and let the LCF students use it.  We have a Digital Learning Lab where the things we’re working with at any particular time get housed, and we run open sessions where students can elect to come in, get their hands on cutting edge technology, and participate in the kind of knowledge exchange that’s vital for getting the next generation ready to go into industry.

The Interline: And what’s the appetite like among students for technology?  You’ve said that being prepared to interact with technology is pretty much a prerequisite for any graduate going into fashion today, but do younger people see it as something they’re going to be forced to use, or something they’re actively excited about?

Matthew Drinkwater: The interest is there for certain.  Just a couple of years ago we ran something called the “Future of Fashion Incubator,” which had Microsoft providing us with hardware and software so students could look at artificial intelligence, wearable computing, mixed reality, and other key areas.  We saw really big take-up of that.

But there’s a broader question of how we embed what we do into the curriculum.  Today all the hands-on opportunities we offer are elective.  And anything we do that draws publicity, or has resources put behind it by a company like Microsoft, gets quickly over-subscribed.  Our new AI course, for instance, only had space for 20 students, and we sold out in a day – with a waiting list of 80 people.  So there’s more demand there than we can supply, for sure.

On the positive side, though, it’s joyous to get a hundred percent attendance on our courses.  That’s really unusual.  And what’s even better is that our AI course had a hundred percent female attendance, which is wonderful to see for a technology-focused course.

The Interline: That’s interesting, because fashion tends to skew quite female even at the executive level, but the tech side has been quite dominated by men.

Matthew Drinkwater: What’s also interesting is that we’re seeing skillsets among our students that we never would have seen five or six years ago.  Today, we have students turn up for fashion degrees who also have the ability to code in Unity and Unreal [the two most popular freely-available game engines – Editor] and while that’s not commonplace, it’s happening more and more often.  So we have an emerging workforce where hybrid skills are becoming more prevalent, and that’s a genuinely exciting thing for the industry.

The Interline: The Unity and Unreal thing is particularly interesting, because up until now there’s always been a barrier in the way for people who want to learn digital design or 3D visualisation.  Having to learn to use AutoCAD or a 3D solution – both of which are typically paid for – stood in the way of young people developing the kind of hybrid skillsets you’re talking about.  But the thing with those engines is that they’re incredibly robust and enterprise-ready, as well as being free and relatively user-friendly for people who understand game engine and real-time rendering principles.

Matthew Drinkwater: Definitely.  I think that as we begin to foster these skills, and expose up-and-coming talent to the kinds of technologies that are coming in the next three to five years, that’s how fashion is going to develop a more creative relationship with technology in general.

So much of what we see in fashion today – particularly luxury – is that conveyor belt mentality, where companies keep doing what they’ve always done, and trying to do it faster.  I believe bringing in these new hybrid skillsets is how companies are going to discover new business models and genuinely shift in new directions.

The Interline: From that point of view, are we going to see a shift in job roles in the near future?  Take machine learning, which is a really technical discipline.  Are we more likely to see technology vendors figuring out how to make the power of computer vision, deep learning, and other technologies totally intuitive to use, or are we going to see brands and retailers recruiting the kind of tech expertise they need to build models and neural nets in-house?

Matthew Drinkwater:  There was a lot of publicity recently, around this time last year, when Marks & Spencer hired their first data scientist.  Chief Data Science Officer, I think the title was.  But I think you’re going to see a lot more of that when you look at the high street, where it’s virtually impossible to buy anything at full price.

[NB: This interview was conducted before the coronavirus outbreak led to the shuttering of most retail premises and the UK entered its lockdown period – Editor.]

The reason for that is that big retailers have been operating on gut instinct for so long, with merchandisers looking through pages upon pages of sales results from multiple stores, and trying to build a story to tell the buyer, who would then make their own intuition-driven assumption of what might be right for the market.

I think it’s inevitable that those job roles are going to change.  As we develop the ability to pull more real, quantifiable data points into those decision-making processes, computer science is all but guaranteed to become a much more prevalent part of the buying and merchandising side of retail at the very least.

The Interline: And in your experience, what’s the perception of Artificial Intelligence (which we’re using as catch-all term for a lot of different technologies) in the industry and among students?

Matthew Drinkwater: Look at industry, I think you’d probably see a big difference between traditional retail and some of the interesting startup players who began their lives in eCommerce.  Think of companies like Farfetch, which invested a lot in building a computer science team, and Stitch Fix, which used data to build its entire business model.

Talk to more traditional retailers, and I think you’d see very similar attitudes to the rest of the general public: that AI is steering us all towards a utopia or a dystopia.  And it’s unfortunate, but AI is saddled with such a legacy from fiction that it’s going to be a huge task to get people to revisit the assumptions they’ve inherited from books, films, and videogames, and to see the actual, practical applications here and now.  And even today you see so many industry articles about AI that start with an image of The Terminator, which frustrates me no end.

The Interline: If it helps, our sister publication, WhichPLM, did a 100+ page deep dive on AI a couple of years ago, and about 20 pages of that was devoted to helping to explain the difference between general and narrow intelligence, and I don’t think we used a single image of a killer robot.

In our experience of speaking with brands and retailers, though, the biggest barrier to really understanding AI and machine learning is how invisible it is – despite being everywhere in our daily lives.  From tagging faces in Google Photos to recommending songs on Spotify, machine learning is already active everywhere, but people don’t realise it.

Do you think AI has the potential to follow a similar path in fashion retail, where it’s active but invisible? And where would the most likely places be for that to happen?

Matthew Drinkwater: The area machine learning is currently the most active and the most recognisable is in forecasting.  But search, recommendation, and personalisation are the areas where the most work is being done at the moment.

The Interline: What do you see as the most interesting applications of machine learning coming over that three-year horizon you talked about?

Matthew Drinkwater: Looking that far ahead pushes us into some really fascinating creative applications of AI.  Visualisation is one of those areas, and we’re also looking at generative models as well – generative design – as well as more in the way of computer vision and automated image manipulation, so image transfer, inpainting and such.

These are the areas where we can see a young designer, in the near future, wanting to use these technologies to do something genuinely new.  That, to me, is where AI gets super interesting.

Take the coronavirus outbreak.  Fashion shows that typically would have happened in front of a crowd are now being played out behind closed doors, or cancelled entirely.  So I’m being asked a lot of questions about how we could rethink that experience virtually, and we’re starting to look at what completely digitally-generated shows might look like – experiences that could be viewed from wherever you are in the world, and that offer an opportunity for brands to express themselves in new ways.

The Interline: We also need to consider what this all might mean for the audience of runway shows.  As you know, consumers aren’t the target for fashion shows, and it’s looking increasingly unlikely that the same numbers of people are going to be willing to travel to shows in the future.  So whether you’re showing physical products or something totally generated or synthesised, doing fashion shows virtually removes a big barrier there, at the same time as opening up an incredibly new marketing and market-testing tool by inviting consumers into a conversation they haven’t historically been part of.

Matthew Drinkwater: Definitely.  That’s going to be a really exciting area to explore.  We’ve also talked a bit, at the Innovation Agency, about ‘deepfakes’, and I know some of the students we had last year explored using that technology to put different clothes on people’s bodies.  And you don’t need to go very far from there to see how a thirty-second facial scan could allow someone to not just view a virtual collection but then try it on, instantaneously, afterwards.

We’re very soon going to reach the stage where that’s not just credible, but believable-looking, and there are guaranteed to be lots of other opportunities to use that kind of content elsewhere.

The Interline: On the plus side, that means neither of us is likely to be short of work any time soon!  Thanks for your time.


To explore more of the Fashion Innovation Agency’s work, visit http://www.fialondon.com/

You May Also Like