What will advertising and media look like in a future dominated by algorithms?

At the MIT Media Lab, a frontier of cutting-edge technology, diverse and advanced research is conducted. Within this lab, a unique project supported by the Dalai Lama Center for Ethics exists. It aims to create new ethics by connecting people worldwide with complete strangers for a limited time. Indeed, terms like "ethics" and "humanity" are frequently heard at the Media Lab these days. While these themes might seem far removed from cutting-edge technology at first glance, the rapid evolution of AI and robotics may be what has turned researchers toward such philosophical questions. As the first installment in our series on the MIT Media Lab, we spoke with Associate Professor Kevin Slavin, who leads this project.

[Project Introduction]
"20 Days of Strangers" is a project where participants anonymously connect with a complete stranger somewhere in the world for just 20 days, sharing information about their lives—where they are, what they're doing, and how they feel. Additionally, the "Playful Systems" research team led by Dr. Slavin undertakes numerous projects aimed at visualizing and making enjoyable complex, invisible systems—from financial transactions to search engines—that permeate our world, countering their current state of obscurity.

Social media has shattered the words shared by many people

Ono: The "20 Days of Strangers" project, as evidenced by its support from the Dalai Lama Ethics Center, seems intended to change how we view and feel about others through technology, contributing to a "new ethics," so to speak. However, existing social media has also altered the nature of human relationships. Where do you see the difference?

Slavin: For example, I use Facebook to connect with people I already know, and I use Twitter to connect with people I want to get to know. This isn't unique to me; it's how social media is used. The problem is that, as a result, social media may inadvertently be narrowing our world. The internet was originally based on the principle of connecting the world and creating a broader one, but in reality, the opposite is happening.

That's why I wondered: what if we created software that connects people who are complete strangers, people who would never have met otherwise? Could we build genuine human connections, where strangers connect and share feelings, all while maintaining privacy? That thought sparked the project "20 Days of Strangers."

That was the starting point, but it ultimately made me clearly recognize the need to create technology that makes people feel they belong to a larger group unrelated to themselves. Furthermore, I realized that direct communication isn't always essential for us to understand something about each other. Even without conversations or exchanges, we can grasp the fact that beyond ourselves, there are countless other lives in the world.

Ono: Many experts in Japan point out that the current internet is actually making the world smaller. Is this an inevitable problem inherent to the internet?

Slavin: I believe so. This problem inherent in social media became starkly evident during the recent U.S. presidential election. There, the concept of widely shared language collapsed. A key reason is social media's tendency to connect people only with those who think similarly. During the heyday of 20th-century mass media, broadcasters had to find words that would resonate with as many people as possible. But social media does the opposite; it gathers only those with similar views. It possesses a strange gravitational pull that pulls people away from the center.

Ono: Is there a way to address these problems of the internet?

Slavin: Seeking out like-minded peers is a fundamental human instinct, and the internet is simply very good at facilitating that. Rather, we should see the 20th century as an exceptional era that went against that instinct.

Ono: I've heard similar points from Japanese experts—that the 20th century, when mass media held immense power, was a unique era, and the internet is actually restoring things to how they were before.

Slavin: That's probably true, and it's neither good nor bad. In principle, the internet should support various forms of communication, and this has many benefits. Indeed, positive ideas never seen before are constantly emerging online.

What we've lost is the experience of many people sharing the same time and perspective. When I was a child, there was a Thursday 8 PM TV program watched by 60 million people simultaneously. By Friday, everyone knew its content—meaning everyone shared the same perspective on the world. Indeed, I believe it was incredibly important that even someone living alone could have that experience. That was the essence of broadcasting. But that no longer exists. The question is whether the internet can restore it.

Will future advertising move away from screens and return to things, experiences, and "voice"?

Ono: You also have experience working at an advertising agency. What are your thoughts on the relationship between advertising and digital technology?

Slavin: The biggest challenge for advertising today is the ability to deliver ads with high precision. This has long been advertising's dream. Advertisers only want to show ads to people for whom the exposure matters. But advertising's function isn't just that; it's also important that others know about it. Meaning, because others care about that brand, I start to care too. If an ad truly becomes just for me, it actually loses its value, right?

Ono: What will advertising look like in the future, driven by technological advancements?

Slavin: I think OOH (Out-of-Home) advertising, real-world spaces, and physical objects will become incredibly important. Right now, we're constantly moving from screen to screen, but the importance of screens will diminish. I believe advertising will shift back toward physical objects and experiences.

Ono: Are you thinking of things like wearable devices or virtual reality (VR)?

Slavin: Personally, I don't see VR becoming a mass phenomenon like mobile phones. It will play a meaningful role in entertainment, but it won't become a fundamental experience provider like today's mobiles or PCs.

Conversely, technologies that are "invisible" rather than visually "visible" will become crucial. Specifically, the realistic long-term scenario is people interacting with devices primarily through "voice." Voice becoming the main interaction method poses a significant challenge for advertising.

Furthermore, as AI advances, people in the future world will no longer search around for information; they will simply obtain the correct answers provided by AI. That's fine in itself, but it also raises the question of what advertising can actually do in that context.

We need principles that can counter the accelerating "culture of algorithms"

Ono: Regarding that AI, as you've mentioned in other talks, AI algorithms are encroaching on the world and carry risks, like the flash crash* that occurred in the financial industry. How does this look from an advertising perspective?

Slavin: The concern with advertising isn't something like flash crashes, but rather what's called ad fraud**. It's similar to what happens in the highly automated financial industry. Essentially, you have AI instructed to "reach humans with ads" on one side, and AI instructed to "pretend to be human" on the other. If I were in the position of delivering ads, I'd want machines that can tell if they're reaching humans. Conversely, if I were an illegal publisher, I'd want to pretend to be human. These two sides compete, each becoming increasingly sophisticated, so it's a never-ending cat-and-mouse game.

Ono: You also noted the incident where a book normally priced around ¥10,000 on Amazon was assigned a ¥2.6 billion price tag due to an algorithm runaway. The danger of AI and algorithms slipping beyond human control exists across various fields, including advertising, right?

Slavin: That danger won't manifest like in the Terminator movies. It will happen by altering how people perceive the world. The media that shapes how people think and feel about the world is already driven more by algorithms than by editors. This is based on the economic principle of optimization, which reinforces whatever ideas it encounters. This can clash with social ethics, potentially exposing people to unwanted ideas and media.

Ono: Algorithms on e-commerce sites are built from an economic perspective. However, some experts point out that a different principle is needed from the perspective of creating a rich culture.

Slavin: Absolutely. While some claim that e-commerce sites, with their vast data, truly understand people, I believe it's actually writers and editors who possess the deepest understanding of human nature.

Humans have two conflicting desires: the urge to belong to a group and, conversely, the urge to differentiate themselves from others. The culture of e-commerce sites and algorithms caters very well to the urge to belong, but it doesn't cater to the urge to differentiate. Humans are constantly seeking something new, and that cannot be incorporated into optimization derived from accumulated behavioral data.

Ono: To address such issues, is human intervention still necessary at some stage in the algorithmic process? There's also the idea that in chess, the strongest teams are those where humans and machines collaborate.

Slavin: Actually, currently, the strongest combination in the chess world is said to be when two human players team up without a machine, discussing and strategizing together as a team. The problem with current computers is that they are designed to operate independently. The key going forward will be how to create computers that can cooperate with each other.

Ono: What about the division of labor between machines and humans? In reality, drawing the line isn't easy, is it?

Slavin: There's a good example of this problem. Last year, a massive blizzard was forecast for the northeastern United States. It was predicted to be the worst blizzard in history, so New York City shut down all municipal facilities, including the subway. The downtown area became deserted, causing significant economic losses. But in reality, that blizzard never came.

The computer had 17 prediction models, but only one predicted the blizzard. However, that model had always been highly accurate. So, even though the other 16 models predicted no blizzard, they were deemed less reliable, and the computer output only that single result. But if someone had been there to see all 17 models' results, they would have thought something was wrong. I believe this incident will be a major lesson for the future.

---

*Flash Crash: A sudden, massive plunge in the stock market. Automated and mechanized financial trading systems have been cited as a cause.

*Ad Fraud: An illegal advertising practice where computers, not humans, mimic human behavior to generate ad impressions and clicks, thereby inflating advertising costs.