Subscribe to Here’s the Deal, our politics
newsletter for analysis you won’t find anywhere else.
Thank you. Please check your inbox to confirm.
Watch Part 2
How big data became a silent colleague for artists and designers
Big data is disrupting nearly every aspect of modern life. Artificial intelligence, which involves machines learning, analyzing and acting upon enormous sets of data, is transforming industries and eliminating certain jobs. But that data can also be used to appeal more directly to what customers want. Special correspondent and Washington Post columnist Catherine Rampell reports.
One of the fundamental economic shifts of our time is the way that big data is disrupting commerce and everyday life.
Artificial intelligence, which involves machines learning, analyzing and using enormous sets of data, is expected to have an ever-wider impact, transforming industries and eliminating some jobs.
That data also can be used to appeal more directly to what customers want, including in creative industries.
Special correspondent and Washington Post columnist Catherine Rampell has the first of two stories on that for our segment Making Sense.
This is called Petit LU. It's an interpretation of a little French butter cookie.
Oh, it smells so good.
And that's it.
I don't know that I want to smell like this, but I want to eat this.
Right. But you want a little of that.
David Apel is a master perfumer at Symrise, one of the world's largest fragrance and flavor design companies.
It's also on the creative cutting edge, harnessing the power of big data to make artistic decisions, part of the second digital disruption, as legal scholars Christopher Sprigman And Kal Raustiala call it in a new paper.
So, the first digital disruption was really about the ability to distribute digitally.
Legally, or not, given the piracy on file-sharing platforms like Napster. Creative industries had to adapt to the new ways people were getting their music and movies,
The second one, which is happening right now, is really about data and not just distributions. It's streamed out to you, but then, in turn, the company is receiving data from you.
So this is all about targeting consumption to the consumer. What do you want? If you go a little bit further, you get into actually investing in content.
Companies like Netflix pay close attention to the details of what people watch, down to when they hit pause. User data is mined, analyzed, and then used to create new products. The classic example? "House of Cards."
Netflix was willing to green-light that series without producing a pilot. They were confident enough in what they thought they knew that they spent a huge amount of money kind of sight unseen.
They identified that people like this particular original British version. They like David Fincher. They like Kevin Spacey. And so we're going to put those things together, and we know there's an audience for that kind of combination.
That turned out, at least until Kevin Spacey ran into some trouble, to be a very successful show.
Even savvier at using big data to create content? The porn industry.
For a few reasons, they probably use data more than other companies do in their creative efforts, one, because, typically, adult films are short. People watch a lot of them. They get a lot of data. They have massive viewership. And they're cheap to make.
Well, I have been approaching it scientifically as of late. Why not, right?
Better porno through science.
Of course, even in pre-Internet days, the porn industry was studying customer tastes.
The professors cite the peep show operator in HBO's "The Deuce," set in the '70s.
Last couple of months, I have installed the quarters in each machine separately. Now I know what film I'm running in each machine, and I never mix it up before we do the weight. It's starting to give me a real sense of what stuff is bringing in the most quarters.
What we were living in the past, you could think of as the data bronze age, right? So, this has now become a gusher of data that's much cheaper to gather and much cheaper to analyze.
Now, most academics shy away from studying porn. But these two used it as a case study, because it's been so innovative, down to using data to write scripts.
So what are examples of the kinds of creative choices, let's say, an adult film company might make based on an analysis of data? Is it like, if there's a plumber at the door or a pizza delivery guy, or…
Absolutely. What does the background look like? What does the room look like? What is the person wearing? Those are all things that can be altered and manipulated in different ways to see what's most popular.
Of course, sex sells in any industry. And what smells sexy? Robots are surprisingly good at figuring that out too.
That's a commercial for Egeo, marketed by the Brazilian cosmetics company O Boticario.
It has this new hot milk, honey kind of note.
The fragrance was designed by Symrise, with some help.
And there you go, the world's first ever A.I. fragrance.
That's thanks to a partnership with IBM Research.
This is "Jeopardy."
It's been nearly a decade since Watson won "Jeopardy."
Now we come to Watson. "Who is Bram Stoker?"
(CHEERING AND APPLAUSE)
Since then, IBM's artificial intelligence technology has gone to work in medicine, science, manufacturing, business. But, recently, it's taken on more artistic tasks.
For a long time, people have been asking whether computers can be intelligent, which is kind of a hard philosophical question to answer. So our group decided that we were going to tackle a different problem, which was creativity.
But, says IBM's Richard Goodwin, a lot of people were already working on android art or A.I. music.
So, our original idea was to actually build a robot that would go on a cooking show and actually, you know, compete with the other chefs. But we then realized it would be like 10 years before we could get the robot to, like, chop up the carrots.
The dexterity is the problem.
And so the idea then was, well, what really wins the show is having novel dishes that taste good.
Enter chef Watson, and then a collaboration with McCormick spices.
From flavors, a short leap to fragrances, a labor-intensive industry seemingly ripe for disruption.
The work of a perfumer is mostly frustrations.
Designing a new fragrance has historically required a lot of human capital. It takes four years of schooling to become a junior perfumer, years more to become a master like David Apel, who has four decades under his belt.
He's got access to over 1,000 raw ingredients, millions of existing formulas. Infinitely more could be tried.
So if you were tasked with creating a new fragrance, what would your process look like?
People will come. And they give imagery and storylines and visuals, sometimes even music or textures.
And so the first process is, what things in my brain connect to the things they're trying to say? I made a fragrance for Pavarotti back in the day, and I built it around patchouli because patchouli is a material that has like this deep, like, bass kind of resonance, you know, the way I saw his voice.
That flash of inspiration, hear Pavarotti, think patchouli, isn't enough. It can take years of tinkering before a formula becomes aesthetically and commercially successful.
If you think about perfumery, it's an inherently complex business, because every creation is a unique product.
Achim Daub is president of scent at Symrise.
Manufactured out of many ingredients, very often too many for my liking.
So there is an inherent inefficiency. And so one of the aspects of using artificial intelligence is to become faster, become leaner, become more agile.
So, A.I. can cut costs, but can it actually be creative?
I'm going to start with this, which is called beurre, which is…
David Apel gave me a demonstration of how the software, called Philyra, works. He starts with two existing elements.
And the first thing that Philyra will do is, she will put them together 50/50.
And then cranks up the creativity, literally.
So, sliding this creativity scale up to 10, I just want her to play.
In 90 seconds, 1,000 possible candidates…
And here we go.
… 12 top contenders.
I like the suggestions that she's given me, and I just push a button, and it's sent to the lab and compounded, and I can smell and evaluate.
How has the introduction of A.I. changed this world for you?
What's really amazing for Philyra is that she knows not just my style of perfumery, but everybody's style of perfumery. And she can…
You say she?
She. She's a she, yes.
They joke here that she's my girlfriend because I have spent nights and weekends to kind of try to understand the way that this machine seems to think.
It's hard not to anthropomorphize software doing something that seems so fundamentally human.
I want to be immortal.
You know, it's that vain of an occupation, in some sense.
Every perfumer wants to create that uniqueness, that sort of — that magic of something that you haven't seen before in a perfume. And that's what Philyra has done for me.
Could we get to a point where robots take the jobs of artists? In a blind smell test for that Brazilian client, Philyra's fragrance did beat out a scent created by David Apel.
They're definitely not going to take all the jobs. They're going to help people who are creators give the public what they want.
Getting better at giving the public what they want raises some legal and economic questions, though, questions we will explore in our next segment.
For the "PBS NewsHour," I'm Catherine Rampell in New York.
A lot to worry about.
Watch the Full Episode
Support Provided By: