Innovation Monitor: Exploring the intersection of Art & AI

This week, we’re learning about artistic uses of AI.

Welcome to the 29th edition of the NYCML Innovation Monitor.

AI is in the eye of the beholder
This week, we’ll look at how AI is influencing — of all worlds — the art world. As creativity and the creation of art continue to incorporate more blends between human and machine processes, it raises questions on authorship, among others. We asked ourselves: As we’re teaching machines to visually, linguistically, computationally process our world…

…can we teach computers to be creative? What are the technologies underpinning new creative expression? We’ll take a closer look at generative adversarial networks (GAN), robot-generated paintings, and wave to Lil Miquela, the avatar influencer that brings this blend of art and AI into the commercial spotlight.

We’ll also get up to date on Google’s recent first-ever chief health officer hire, learn about VR’s potential as a teaching tool for surgeons with surprising outcomes, and wholeheartedly embrace metrics even in the face of criticism around the “tyranny of metrics.”

We hope you’ve been enjoying this newsletter and would love any feedback (erica@nycmedialab.org), especially in these early stages. Thank you again for reading!

Best,
Erica Matsumoto
NYC Media Lab

AI ART’S ASCENSION

Last year, in a sale that shocked the art world, storied auction house Christie’s sold a piece of AI-generated art for a whopping $432,500 (for context, this exceeded the combined value of an Andy Warhol print and Roy Lichtenstein work next to it). This was a major moment for AI art, and marked something of a turning point for AI’s use in “serious,” investment-level art.

Now, a year on, AI art has had more big, notable moments:

• “Faceless Portraits Transcending Time”: This exhibition at the HG Contemporary gallery in Chelsea earlier this year was a collaboration between AICAN (an AI) and its creator, Dr.Ahmed Elgammal. It was the first solo gallery exhibit devoted to an AI artist.
• “Eternal Golden Braid”: In March 2019, Oxford mathematician Marcus du Sautoy, harpsichordist Mahan Esfahani, and composer Robert Thomas held an unusual event in which attendees listened to music partly composed by Bach and partly by AI, and tried to guess which parts were which by holding up two-sided cards with a blue human face on one side and a red robot face on the other.

How does it all work, though? Most AI art uses a machine learning architecture called the generative adversarial network, or GAN. This works by pitting two neural networks against each other in a two-part model:

• Generator: attempts to create realistic, but not counterfeit, data (generally, images)
• Discriminator: receives images from the Generator, along with real images from the training dataset, and attempts to differentiate between real versus generated images

Every time the discriminator successfully rejects the generator’s output, the generator tries again to fool the discriminator. Martin Giles compares the interaction to “the back-and-forth between a picture forger and an art detective who repeatedly try to outwit one another.”

To get a better handle on how this technology works, you can try your hand at using either of the following technologies, both of which were buzzy earlier this year:

• Ganbreeder: uses breeding and sharing to explore high complexity spaces, allowing users to generate images through generating “child” images and “crossbreeding” generated images with other Ganbreeder images
• GauGAN: a simple exploratory tool that allows both beginner and expert artists to generate AI art, and a nod, perhaps, to painter Paul Gauguin

Of course, GANs can be used in potentially dangerous or questionable ways, such as to create deepfakes to spread misinformation or to forge fingerprints and other biometric data. The potential to misuse GAN for creating fake images is real; and researchers have already demonstrated it with regard to human faces, cats, anime, and more (if you’re looking to fall into an AI hole, the challenge of distinguishing between real versus fake people is a great way to lose hours of your life).

AI IS HERE TO STAY

AI has enthralled the public thanks mainly to the media’s misconceptions and exaggerations, so it’s not surprising that AI art is becoming ever more common.

Despite his criticism of Obvious last year, Ahmed Elgammal called AI-produced pieces a form of conceptual art, “a form that dates back to the 1960s, in which the idea behind the work and the process is more important than the outcome.”

Today, a year after Obvious’ breakthrough sale, “AI artists” are raising the bar. Ai.Da, considered the first ultra-realistic drawing robot artist, can draw, paint, and even sculpt (in short, she’s more multi-talented than many human artists).

Source: Singularity Hub

Her first exhibition, Unsecured Futures, was showcased at Oxford University in July.

Artists who are worried the machines are coming for their jobs needn’t worry, though: art curator Fabian Offert argues that AI art is neither an artistic revolution nor an existential threat to human artists. Instead, he says, it’s merely “the continuation of a process we already see today: the slow and steady recuperation of explicit ‘AI art’ by contemporary art.” In other words, machine learning is merely a new set of tools in artists’ toolboxes (a phrase we can maybe apply to much of the AI / ML disruption conversation).

HBR Research: How Virtual Reality Can Help Train Surgeons

Although advancements in medical devices and surgical techniques promise to improve the tools surgeons have for saving lives, the system for training and assessing surgeons themselves has lagged behind the pace of innovation. This leaves some doctors unprepared to perform complex surgeries and puts some patients at risk. A recent clinical validation study at UCLA’s David Geffen School of Medicine using VR training could change this. The study found that VR training improved participants’ overall surgical performance by 230% compared with traditional training methods. 5 min read Read More

Google appoints former Obama health official Karen DeSalvo to new chief health officer role

In another indication of its aggressive pursuit of the health market, Google has hired ex-Obama administration official Karen DeSalvo to serve as its first-ever chief health officer. She joins former FDA Commissioner Robert M. Califf and former Geisinger CEO David Feinberg. Together, these three hires significantly bolster Google’s life sciences and health industry expertise as Alphabet makes broad investments in the health industry, including in new drugs and devices and cloud computing for life sciences companies. 2 min read Read More

Don’t Let Metrics Critics Undermine Your Business

Economic historian Jerry Z. Muller is credited with the phrase the “tyranny of metrics.” “The problem is not measurement,” Muller declares, “but excessive measurement and inappropriate measurement — not metrics, but metric fixation.” Even one of the more recent Harvard Business Review cover stories was “Don’t Let Metrics Undermine Your Business.”

Michael Schrage from the MIT Initiative on the Digital Economy, in this MIT Sloan piece, encourages us to get back on #TeamMetrics, and explains how measurement works best with the right strategy and incentives in place.

12 min read
Read More

This Week in Business History

October 26, 1999: The Wall Street Journal announces that Intel and Microsoft will be added to the Dow Jones Industrial Average

These two corporations are the first from the NASDAQ to be listed on the Dow. In an indication of the entire economy’s shift away from “smokestack” enterprises, they replace Chevron and Goodyear. It’s also announced that Sears Roebuck and Union Carbide will be replaced on the Dow by Home Depot and SBC Communications (maker of the Baby Bell telephone).

This email was sent to <<Email Address>>
why did I get this? unsubscribe from this list update subscription preferences
NYC Media Lab · 2 MetroTech Center · Brooklyn, New York 11201 · USA