[Feature: Social Media and Society] Roundtable Discussion: The Accelerating Attention Economy and the Future of Social Media

Today, I would like to discuss the theme of "Social Media and Society" with all of you.

Currently, platforms such as X, YouTube, TikTok, and Instagram have become indispensable for information distribution in Japanese society, and their influence continues to grow. However, various issues have also emerged, such as the spread of fake/false information, conspiracy theories, and defamation. Within this context, I hope we can have a constructive discussion today toward "making digital spaces healthier" and "forming better information infrastructure and public spaces."

To begin, I would like to ask everyone to tell us about your current work and what concerns you most about social media (SNS) right now. Thank you.

I am a reporter in the Science and Culture Division at NHK. This is my 10th year, and since I am assigned to the IT team, I cover social media and cybersecurity.

Currently, the themes I cover most frequently are related to advertising. I often report on issues such as malicious advertisements slipping into social media, impersonation ads, fake brand-name ads, ads leading to malicious seminars, and ads containing extreme sexual expressions that are inappropriate for children.

Other areas of interest include the flood of clickbait content designed to attract advertising revenue and the impact that opaque algorithms are having on users. I hope to cover these topics in the future as well.

Additionally, as for current topics, I am broadly covering issues such as the promotion of polarization by social media and the overflow of fake/false information.

I joined the former Yahoo Japan Corporation in 2000 and have been in this industry for 25 years. Initially, I was in the Legal and Policy Planning Department, primarily involved in copyright work and examining the state of the copyright system. In 2017, I moved my primary duties to Yahoo! News, where I work on matters related to the health of the information space. I have been involved in what you might call rule-making within Yahoo! News, such as establishing guidelines for articles distributed by media companies, making the "Yahoo! News Editorial Policy" transparent to explain how we operate the service, and consulting with experts to hear their views.

Regarding social media, a point that has concerned me recently—especially looking at the recent House of Councillors election—is whether information of high public interest is reaching people, particularly the youth. This is often called "prominence," and I believe we must ensure that information of high public interest, such as elections and disasters, as well as high-quality daily news, reaches people effectively.

I first joined a system engineering company, then changed jobs to an advertising-related company, which led to my current position. I have been involved in digital advertising for about 15 years.

Regarding what concerns me, I recently returned to my hometown in Kagoshima, and my 72-year-old mother, who runs a Juku, was running digital ads. When she asked me, "What do you think of this CTR (click-through rate)?" I realized how far it has permeated. Digital marketing is becoming common even for owners of small local Juku. What general users currently think and how they perceive digital marketing is the point I am most interested in right now.

Another point of concern is advertisements attached to hate comments. This is a complex issue involving the unit price for advertisers, the monetization of the publishing media, and the billing methods of the ads themselves, so it is not straightforward. I hope we can consider solutions through technology and the knowledge and awareness of users.

In my 20s, I was a national public servant in Kasumigaseki, and after entering my 30s, I knocked on the door of academia. Before I knew it, I was focusing on social media research, primarily using a data science approach. I collect digital data from YouTube and X and write papers on people's information behavior.

Recently, I have been studying the provisions regarding data portability in Article 6, Paragraph 9 of the EU's Digital Markets Act (DMA). Essentially, this says that network data generated by users and end-users on social media must be made portable. I believe the goal is to reduce the powerful network effects monopolized by platforms and aim for a future where people can freely choose their social media.

I am often asked by the media about how SNS should be regulated in the future, but regulations based on the content of expression, like those currently being considered, are very problematic in relation to the Constitution. If we are to do anything, I think one way out is to arrange rules in a direction where people who dislike X, for example, can choose the information space they like, such as Bluesky or Mastodon.

Thank you, everyone. You have raised various interests and issues, but first, I would like to discuss the "merits and demerits of social media as an interactive medium."

First, regarding the "merits" of social media, I think the fact that everyone has gained the opportunity to become a sender of information is very significant. The media environment before that was basically centered on mass media, with a clear distinction between "senders" and "receivers." With the spread of social media, the barrier to entry for becoming a sender has dropped considerably. Now, anyone can become "like" a critic or a journalist.

On the other hand, it is also said that the influence of so-called legacy media, which previously monopolized the media for information dissemination, may be declining.

Next, regarding the "demerits," there is harmful information such as fake/false information, conspiracy theories, and hate speech. Furthermore, there is the issue of illegal defamation, which is also covered by the Act on Improving Information Distribution Platforms (Information Platform Act). In other words, one point is how to think about the fact that user-generated content causes harmful effects because it has become a two-way medium.

Then there is the talk of advertising. There are cases where the ads themselves are harmful, such as so-called scam ads or impersonation ads. However, even if there is no problem with the ad itself, there is the issue that by attaching ads, they become a source of revenue for those who disseminate fake/false information, conspiracy theories, or hate content, providing them with economic incentives.

First, Ms. Kinukawa, since becoming a reporter and following these social media-related issues, have there been any events that felt like a turning point?

I am of the generation that got their first smartphone upon entering university, and at that time, Twitter was still quite idyllic. When I started working in 2016, there was a trend even among local governments to actively use social media. This is just my personal feeling, but I feel that the presence of social media grew significantly around 2020, before and after the COVID-19 pandemic.

Originally, I spent a long time covering police cases at the Wakayama and Kyoto bureaus. Previously, cyber-related topics were mostly about unauthorized access by young people or youth getting involved in crimes. There was a clear divide: social media and the internet were for young people or experts, while the elderly were more likely to be victims of scams over the phone. I think that has gradually blended, and now people of all ages are becoming victims via social media.

I don't know the exact turning point, but I have the impression that when the movement to connect online emerged during the pandemic, social media became more familiar, and crimes and troubles also spread into the SNS space.

I see. Mr. Imako, how do you feel about this?

This might deviate slightly from social media, but I would like to talk about Yahoo! Kome (the comment section of Yahoo! News). Yahoo! Kome was established in 2007 and developed as a place where various social issues are actively discussed. On the other hand, there were issues such as defamation, so Yahoo! News has discussed this intensively and taken proactive measures. We are moving forward with not only human patrols but also using AI for automatic deletion. Additionally, we have experts from various fields participate as official commentators and display their comments at the top. As a result, I feel these measures are proving successful.

Since fiscal 2020, we have voluntarily published a Media Transparency Report every year, disclosing the systems and measures we have in place for posting-type services, how many posts were made, and how many of those were deleted.

Looking at Yahoo! JAPAN's posting-type services as a whole, compared to fiscal 2021, the number of posts themselves has not decreased in fiscal 2024, but the number of deleted posts has decreased to about one-fourth.

Thank you. Mr. Magome, have there been any major changes in the field of digital advertising over the last 10 years or so? I imagine the acceleration of the attention economy is one such point.

When I entered this industry, ad space was sold as fixed slots rather than programmatically, but the landscape has completely changed with the spread of programmatic advertising and SNS advertising.

When we first started selling SNS ads, targeting could be done more flexibly than it is now. Because the targeting accuracy was high, the resulting effects were also very high. After that, regulations on personal information and other matters became stricter, leading to the current state.

Current SNS advertising is just as effective as other programmatic advertising. I think this is because the platform algorithms are excellent and the targeting accuracy is superior. Therefore, they are selling very well and continue to grow. I believe this trend is irreversible and will not go back.

What Mr. Magome just said is probably very relevant to the fundamental part of the attention economy. We also need to keep in mind that the current trend is "irreversible."

Recently, we have seen more data science insights regarding social media in newspapers and other media. The most recent House of Councillors election is a prime example. Since social media is also a tool that supports democracy, I think the information consumption behavior of users and voters during elections is important. Mr. Tanihara, are there any challenges or problems you saw from social media recently?

During the July House of Councillors election, we conducted a fairly thorough survey before and after the election and also tracked digital data from YouTube. Broadly speaking, the results showed a complete generational divide regarding the relationship between media contact and supported political parties.

With the age of 50 as a boundary, people in their 60s and 70s get information from TV, newspapers, and election bulletins, and this group supports the LDP, the CDP, and Komeito. Those in their 40s and younger primarily get information from social media outlets like YouTube and X, and they support the Democratic Party for the People and Sanseito. This is a very clean divide; it wasn't a split between right and left, but rather a very clear split based on media contact and age group in this House of Councillors election.

This time, the overall voter turnout also increased by about 6% compared to the last time. It appears that younger generations, who previously did not obtain political information, are now getting it through social media, especially YouTube. Of course, information on YouTube is a mixed bag, but in whatever form it takes, I think it is basically a good thing that people get election information and voter turnout increases.

However, at the same time, it is said that about 40-some percent of YouTube videos related to the House of Councillors election are "clipped" videos by third parties. Official videos from political parties account for only 10% to 20% of the total. The rest are videos that TV stations and newspaper companies have put on YouTube or videos by web media.

I think it is basically not a bad thing that younger people, who have not been very active in political participation, can see the claims of various parties, including minor ones. I don't think young people are voting haphazardly just because they are influenced by SNS.

I see.

One thing that concerned me is that there are various surveys about the reliability of legacy media versus the internet, but I don't really understand what they are measuring. For example, even on the same TV, the reliability of a daytime variety show and "World Business Satellite" should be completely different.

Therefore, in the survey I conducted, I specifically listed several political and economic channels on YouTube and asked, "Which ones do you watch, and how much do you trust them?" When asked broadly, TV and newspapers have higher reliability than YouTube or X. However, between "PIVOT" and TV, "PIVOT" has a higher reliability by a fair margin. So, while understanding that the internet in general is a mixed bag, I think people feel that the channels they watch are more reliable than TV.

To add one more thing, the biggest difference between Japan and the West is that in the 2024 and 2025 elections in the West, the main medium has shifted to short videos. Specifically, TikTok. Italian Prime Minister Meloni has 2.5 million followers, and Jordan Bardella, the leader of the National Rally in France, has 2 million followers, which is on a different scale compared to Japan. Japan is still dominated by YouTube. However, since each political party is starting to put effort into TikTok, I feel it is only a matter of time before short videos become mainstream.

In this House of Councillors election as well, each party suddenly started putting effort into TikTok in the final week, and that's where Sanseito and the Democratic Party for the People suddenly gained momentum. There is no evidence that Japanese people like long-form content more than Westerners, so it is highly likely that we will shift in that direction as a natural progression.

Thank you. YouTube and TikTok have quite different recommendation mechanisms via algorithms. Is there anything regarding the relationship between such social media structures—especially the algorithm part—and user information consumption?

The characteristics of the algorithms for YouTube, X, and TikTok, which are mainstream in Japan, are all different. TikTok is the easiest to understand; it is the most closed, or rather, it is all about "optimizing for your tastes." In a sense, it creates such a strong filter bubble that you can even shut out trends.

Also, YouTube was heavily criticized in the 2010s and shifted to a policy of operating quite restrictively. They explicitly state that for harmful misinformation that is on the verge of a policy violation or harsh videos that hurt someone, they won't delete or hide them, but they will reduce their exposure. Furthermore, they say they will give top priority to pushing content provided by authoritative mass media.

Indeed, there are reports that if you follow YouTube's recommendations, you end up watching increasingly moderate content. Therefore, the appearance of highly partisan videos is rather a result of individual choice, which YouTube then learns. Data science has revealed that partisanship is strengthened through a kind of complicity between the algorithm and individual choice.

Regarding X, 50% of the "For You" timeline consists of posts from accounts you don't follow. However, that also ends up recommending only posts that have "gone viral" or right-wing posts.

So, data science has revealed that there are still issues with bias or safety to some extent, but it is not as simple as the "filter bubble is the villain" theory; rather, there is a tendency for content to be displayed through the complicity of individual choices and algorithms that each have their own quirks.

The idea of "complicity" is quite important. Sinan Aral points out in his book "The Hype Machine" that algorithms do not exist in isolation; they continuously learn from user behavior and strengthen "recommendations" through feedback. I think the fact that it is not technological determinism, but rather a mutual relationship between humans and technology, is also related to the merits and demerits of social media.

Ms. Kinukawa, how do you feel about this House of Councillors election from the perspective of a reporter?

This time, NHK introduced the full text of the first speeches of all candidates in the House of Councillors election and reported them with annotations depending on the content. However, I personally feel that I don't yet know how far that actually reached the viewers.

Originally, a mountain of information comes out during an election period. Within that, we work hard in our own way to reach as many people as possible, reporting while trying not to be perceived as "clipping" information. However, I feel we need to verify in the future whether that reached people sufficiently and what the differences were between those it reached and those it didn't.

I see. The possibility that it isn't reaching viewers is a challenge. Mr. Imako, it could be about this House of Councillors election or anything else, but were there any initiatives taken as election measures in the Yahoo! News comment section or the distribution part of Yahoo! News?

From the perspective of fake/false information, we believed it was important to reliably deliver highly trustworthy information in places where it is easy to see, especially during the election period. We posted articles debunking fake/false information on Yahoo! News Topics and distributed articles from content partners that had been fact-checked on the election special page.

What I found wonderful this time was that not only fact-checking organizations but also many companies, such as newspaper companies and TV stations, performed fact-checks. Since they were also distributed to Yahoo! News, we were able to feature them on Yahoo! News Topics, which I think was very good.

In addition, we provided a service called "Political Party Compatibility Diagnosis" that tells users which political party is closest to their own thinking, and it was used by over 3.5 million users. This is a significant increase from the previous (2022) figure, and I felt an increased interest in the election.

In Yahoo! Kome, it's not a situation where fake/false information is rampant, but writing false information about candidates or defamation is prohibited by the Public Offices Election Act and other laws. We alert users and check and respond to problematic posts.

Now, Mr. Magome, digital advertising is basically a marketing story and was thought to have little to do with elections or democracy. However, unfortunately, there are people who spread fake/false information or conspiracy theories to gain revenue. I think there are parts where the digital advertising ecosystem is supporting that.

From the field of digital advertising, how do you view democracy and the health of information?

That's difficult. For now, advertising is very much disliked (laughs). Just as apps called ad blockers, which hide advertisements, frequently appear at the top of sales rankings, the number of users who do not want to see ads is increasing considerably.

There are quite a few people who say things like, "How dare you put up an ad," or that making money is not good. However, I also think that's a bit strange. If ads aren't displayed, the operating company won't make money, so first of all, I want users not to perceive receiving ads so negatively.

Since platforms cannot actually be used for free, if you don't include ads, they will become unsustainable. I think that will lead to a vicious cycle where, as a result, information can no longer be delivered. We have to create a good point of contact between ads and users. Since technology has evolved this far, I think we have to find some kind of solution through technology. Currently, there are no indicators for things like pleasure or discomfort for the user, right?

YouTube is being very clever about that. They operate by lowering the unit price for ads with high user satisfaction and displaying them frequently, while raising the unit price for ads that are skipped a lot and making them appear less often.

Of course, they haven't made the criteria public, but they seem to be making an effort to leave only high-quality ads.

That's true. However, I think there might be slightly different issues regarding whether an ad is "good" or not.

I spoke with major newspaper companies and TV stations during this election, and one newspaper company said that mass media's election coverage this time felt like a different world compared to five years ago. They said each company was very proactive in reporting, including numerical data.

After all, even if everyone becomes a critic, I believe only trained professional journalists can produce facts. Since those facts are circulated on X, YouTube, and TikTok for everyone to comment on, I think the information provided by mass media is reaching people, albeit often in clipped forms.

That's interesting. Ms. Kinukawa, what do you think about that?

Personally, depending on how it's clipped, I sometimes feel like wondering if it's being communicated properly (laughs).

I agree. The problem is that the comment sections for clipped YouTube videos of content dropped by mass media tend to get very heated. Even when I look at things I've appeared in, I'm never praised (laughs). In that sense, I can well understand why it's difficult to put things out on the internet.

And it feels bad even if you close the comments.

When you put mass media content on the internet, even if you close the YouTube comment section, it's common for people to eventually attach that video to X and add their own comments.

When you put mass media content on the internet, even if you close the YouTube comment section, it's common for people to eventually attach that video to X and add their own comments.

On the contrary, proactive so-called "antis" do check things thoroughly, but conversely, I think the challenge for each media outlet is how to deliver information to the "light layer"—those who don't have much interest.

That's true. What I often tell students is that it's not as if newspapers and TV stations never release misinformation. They do issue false reports, and there are times when you wonder why they edited something in such a misleading way.

On the other hand, they have a solid process that includes a certain kind of reporting ethics—an attitude toward putting out facts. The question is whether YouTubers have that. I believe the value of that process must be secured even in the age of social media.

The problem, however, is that the work of gathering facts doesn't make money. Is that public role truly recognized? NHK is a bit different, but I think it's a dilemma between having a public role and being a business entity.

Exactly. Unfortunately, in the current attention economy, "sensational" fakes are more likely to draw attention and easier to monetize.

NHK is a public broadcaster supported by receiving fees, but Kinukawa-san, as a journalist, do you worry about the PV (page views) of your articles?

I do care. I've come to care about how much an article is read and how long people stay on the page. What I feel in that process is that the extreme clickbait headlines often seen online are indeed strong in terms of PV. I think this is closely related to the attention economy, but when I see headlines that miss the essence just to arbitrarily draw interest, it makes me feel uneasy.

Also, in terms of cost, it takes only a moment to send out false information, but verifying facts one by one, interviewing experts, compiling them into an article, checking it, and adding video takes an incredible amount of cost. In that sense, I feel like we are fighting a losing battle.

I believe the system of public media, which should inherently be furthest from the attention economy, is more important now than ever. Yet, even an NHK journalist, who should be the furthest from it, cares about page views. Is that a motivation that is different from revenue?

Motivation is a big part of it, but I also feel that for an article we've spent time and effort on with the cooperation of interviewees, it's only worth it if it reaches people. Back when there was only TV and radio, you might have felt it reached many people just by broadcasting it as a lead news story, but we are no longer in that era. Many people in my generation don't even own a TV. I care about the numbers as a benchmark to judge how many people it reached and to explain that it reached this many people on the web.

It's a kind of magic of "numerical values." It's said that the attention economy originally started with newspapers. There were always issues with tabloid papers and competition for TV ratings.

So, what is the difference between the attention economy of the era when so-called old media was central and the attention economy of today? Magome-san, what do you think?

I think there are just too many interesting things on the internet now (laughs). For example, people in their teens and 20s spend a considerable amount of time on the internet every day.

I think the fact that there are too many interesting things is the primary factor in the competition for time. So, it's not that the media itself is in decline, but rather that the number of competitors has increased enormously. We mentioned "time performance" (taipa) earlier, and I think that's a story about how interesting content is lined up one after another.

Therefore, the time spent thinking is gradually decreasing. If that happens, your consciousness gets pulled away by whatever you just saw, and you become more easily influenced. That's why I think concepts like informational health will become important.

That's a fundamental point. When I talk to students, they say their time "melts away" on social media. I think that describes the current situation well.

In short, "melting" means realizing that several hours have passed or the sun has gone down. I think it's exactly a case of not being able to control one's own time and interests. It's not just a matter of content; by designing how social media "shows" it to us, we are "melting" our time. It's a story about having that kind of "structure."

I'm sure that's right. Very talented people are trying to get users interested in their apps and media, so it's becoming quite difficult for the user side to maintain control.

I'd like to move on to the topic of platform design, which relates to the "structure" we just discussed. The design of algorithms and user interfaces themselves is held by the platform side. In social media, the power of such design works significantly.

I call this the power to form the information environment, but on the other hand, I think this power can be used in many good ways. The story of the diversification model for Yahoo! News comments is exactly that. I'd like you to talk about what LINE Yahoo is doing in terms of platform design.

It might come down to the social responsibility of platformers, but Yahoo! News has long operated its service with the aim of creating a healthy information space.

The "Comment Diversification Model" you just introduced is a mechanism that uses AI to ensure diversity in Yahoo! comments by preventing only similar comments from appearing at the top of the comment list. We introduced it thinking it might also have the effect of mitigating the "echo chamber phenomenon" where specific opinions are amplified. Additionally, in 2024, we began providing a "Comment Revision Model." This is where AI suggests a review of the expression before the post is completed for expressions that viewers might find unpleasant.

Even if it's a gray area regarding guideline violations, posts that might make other users feel unpleasant can ruin the atmosphere if posted in large quantities. When such posts are detected, the "Comment Revision Model" AI activates, highlights the expression, and prompts a review with "Please check," giving the user time to calm down and think. Results show that this has led to a healthier environment. Comparing the four-week average before the introduction of the model with the four weeks after, we confirmed that posts of comments that might be felt as unpleasant decreased by about 24%.

In the social media era, rather than thinking deeply, the brain function called "System 1," which makes hasty decisions, can become stronger, so I thought that giving time to think by having AI request a review might be one solution.

That story clearly shows the importance of considering countermeasures through architecture. I am sometimes told that if we had stricter regulations against online slander through laws, slander would disappear, but I don't think that's possible.

There are issues of freedom of expression, of course, and since legal regulation itself is mainly a reactive response, it inevitably has limits. Therefore, I think one way is to respond with architectural methods like nudges or sludges. The revision model you mentioned might be a kind of sludge in a sense. A mechanism that makes the finger stop before pressing the button.

Unfortunately, however, not all operators actively engage in such initiatives. The challenge is how to move forward when there is little incentive to do it voluntarily.

The mainstream platforms today, both X and Facebook, are centralized servers. In a manner of speaking, the information space is determined at the whim of those at the top. Since Elon Musk took over X, the API (Application Programming Interface) that researchers can access has become paid and limited. Facebook's application process is tedious, and TikTok only allows access in the US and Europe. I don't think that state is healthy.

In contrast, Bluesky and Mastodon are decentralized servers, so users can customize their own information space to a considerable extent. Of course, that has its own criticisms, but I think it's better than a situation where there is only one newspaper.

I believe that if users have the freedom to choose social media with the specifications they like, it will lead to a healthier formation of the information space. At the very least, I think it's exhausting that the entire information space changes at the whim of Elon Musk and Mark Zuckerberg.

I understand. Threads (Meta) is also partnering with the Fediverse (a mechanism for interconnecting and operating independent social networks), and there are proposals for "middleware."

That's right, Threads can now see posts originating from Mastodon. So that's further beyond network portability; in terms of the DMA, it's Article 7 interoperability. It's exactly like how Gmail can interact with other email servers—messages can be interoperable. I think that kind of future is possible.

However, while there are many hurdles like technical challenges and whether you can take third-party information with you, I think we have to somehow break the current oligopoly by a few companies.

However, as you mentioned, Tanihara-san, there is actually quite a bit of criticism regarding decentralized social media and middleware, with some saying that decentralization might make things even worse. Content moderation is a centralized mechanism where the platform sets the rules and manages rule violations like hate, fakes, and slander, using AI to delete them. But if this becomes decentralized, the points of management also become decentralized. There's no guarantee that everyone will do it properly. Nevertheless, I think there is a possibility for a decentralized future for social media as one direction.

Exactly. For example, on Bluesky, if you don't want to see political topics at all, you can label and hide them. Some might call that a filter bubble, but at the very least, it can meet the needs of people who don't want to see slander or false/misinformation.

Above all, the right to choose one's own information space is almost non-existent in social media today, which I think is not good.

In reality, in Japan today, you can't really choose. Text is on X, and since your followees and followers are there, if you're asked if you'll move to Bluesky and start from zero now, you won't. For video, it's YouTube and TikTok. For images, it's Instagram. Since people have already built networks there, they won't move elsewhere and start from scratch.

But in Europe, for example, The Guardian has left X for Bluesky, and many Western researchers have also left X for Bluesky. I think that's a judgment that they are dissatisfied with X's content moderation and that being able to do it themselves on Bluesky is better. But in Japan today, in a state of monopoly, I don't think people feel like moving elsewhere.

In Japan, everyone is indeed staying on X quite a bit. People are creating Bluesky accounts, but the people they followed on X aren't really moving over.

So in a sense, it feels like the opportunity for choice itself, like the freedom to move, doesn't exist in Japan at present.

In places like France, since you can download network data from X, researchers are creating systems where, by uploading that to something like a web app, friends who have accounts on Bluesky can be semi-automatically reconstructed.

Kinukawa-san, from a journalist's perspective, is there anything you feel empirically about the power of platforms?

The main thing is that I often feel it's difficult to interview overseas operators. Even when I do, I can't get detailed information, and sometimes only policies are sent back with a message to read them. In this day and age, the influence of platforms is so large that it could be said they have become a giant "power."

Despite that, a situation has arisen where the actual state of rule operation and the actual algorithms are unknown. As a journalist, I think every day about how to face the fact that platforms have become de facto rule-makers.

Regarding the issue of platform power, a famous legal scholar named Kate Klonick has described them as "The New Governors." They aren't the government itself, but they possess that much power in the realm of content moderation. Professor Tatsuhiko Yamamoto has also pointed out that while the state is compared to the sea monster "Leviathan," platforms are like the land monster "Behemoth."

Tanihara-san's point finds one solution in decentralizing where power is concentrated centrally.

Another direction is how that concentrated power is used. It's quite difficult to control this through democratic laws. The EU has a very strong idea of trying to control it through legal system designs like the DMA, the Digital Services Act, and for personal information, the GDPR (General Data Protection Regulation).

On the other hand, the US is basically libertarian. This has been the case since before Trump; whenever there's an attempt to regulate platforms, freedom of expression inevitably gets involved. So in the US, the discussion turns to antitrust or competition law regulations, but it's doubtful if those are succeeding, and the US has a strong libertarian tendency rather than one of control.

Within that, I'd like to discuss what kind of stance Japan should take going forward. Earlier, Tanihara-san pointed out that legal regulation might be a poor approach, and while I think that's true in a sense, legal studies, including the broadcasting system, have played a part in media design. To move toward a decentralized system including data portability through competition policy, I think there are parts where legal regulation will be necessary. Besides that, do you have any proposals for systems that should exist?

I think something like public authority might be necessary because I'm wondering if something can be done about PV-centrism.

Recently, PV-centrism has permeated all areas, and even certain newspapers have started using clickbait headlines quite a bit. Even though the content is a very solid article, the reason it becomes PV-centric is, I think, because it's an advertising model. If we were to take action, I think that's where it would be.

It really does relate back to the attention economy. I think metrics that emphasize PV and attention are being valued too much.

In that regard, I thought it was a very interesting attempt when Yahoo! News introduced three buttons for evaluating news articles: "Educational," "Easy to understand," and "New perspective." I thought those were metrics that don't emphasize attention.

With a button like "Like" that can be pressed intuitively, you inevitably get pulled into the attention economy. So, as a mechanism to provide a buffer and time to think, we introduced three types of "Article Reaction Buttons" at the bottom of articles: "Educational," "Easy to understand," and "New perspective." We decided that the fact that a user thought and pressed a button means they evaluated it as a valuable article, and we reflected this in payments to each content partner.

How to measure the quality of an article is very difficult, and we want to consider it by consulting with experts. Since there can be negative effects if a platformer decides top-down, I think we need to consider it carefully while listening to various opinions.

The question of who decides the criteria for qualitative evaluation of articles is difficult. In the case of Yahoo! News, since Yahoo is the platform and is in a superior bargaining position relative to media companies, there are issues with unilaterally deciding evaluation metric standards. Who designs something like evaluation criteria that don't emphasize attention, and through what process, is a challenge for the future.

By the way, can anything be done about PV-centrism from the field of advertising?

I think it's difficult. The reason is that the current business structure is largely determined by numbers like PV. I think it's quite hard to suddenly overturn that part.

However, as Imako-san said, I also think things might change as the mechanism itself changes. It's not that clients don't have an awareness of the problem; everyone wants to do something about PV-centrism, but I think they haven't found a breakthrough. If the information space worsens, we won't be able to return good results to our clients, so we don't think it's good at all.

For now, I think it would be good to have something like a user version of ad verification, where users evaluate ads, as a metric.

But sociologically, advertising is supposed to be a media expression that contributes to the culture of the era. However, online advertising doesn't have much of that element. No one talks about online advertising in that way, do they?

Regarding the word "creative" that Magome-san uses in "The Dilemma of the Attention Economy," I felt there was a slight gap with the image I have.

We simply call advertisements "creatives," but it seems ordinary people find that strange.

As you say, we call them creatives because we recognize them as expressive works, but now the elements of a creative are also decomposed in various ways. This relates to PV-centrism, but since we also calculate and create things based on what elements will increase the CTR, I think things have changed quite a bit.

A professor specializing in advertising said that as we move to digital advertising and the attention economy accelerates, artistic elements are being lost more and more.

That's the pitfall of quantification. If you set something like a KPI, you become bound by it.

But even if direct click counts don't show up, there should inherently be invisible effects, like someone buying something at a store because a good image of the ad remained.

That's true.

How to capture the existence of such images feels like a potential breakthrough. Should we make it possible to do so with other metrics, even if it can't be quantified now?

Quantifying the unquantifiable is my job. Even if you can't get a CTR, the final conversion rate (CVR) might be high. I want to manage that through quantification no matter what. Or rather, I think you can't make decisions unless you somehow quantify it.

It's also important to have advertisers properly recognize the negative effects caused by the attention economy and PV-centrism in today's digital media.

How do they feel about the fact that the ads they placed might be running on hate speech videos? While they withdraw ads from Fuji TV, how many Japanese advertisers reacted when Meta relaxed its hate speech rules this January? But it's the advertisers who support the platforms. I think they need to be more aware of the point that their precious advertising costs might be supporting the distribution of harmful information.

Kinukawa-san, how do you see the role of existing media toward platforms in the future?

Regarding platforms, from a user's perspective, I think we have no choice but to honestly and persistently report on them, saying, "Since you have this much influence and public nature, please ensure thorough transparency and safety management." As an individual reporter on the ground, what I can do is steadily report from my own position and disseminate the results.

Monitoring and checks and balances against the government are necessary, but I believe mass media must now take on the role of monitoring platforms that have become the "new rulers."

At the same time, I think news organizations have things they have cherished, such as journalistic ethics. However, it is also necessary for news organizations themselves to update so they are not swallowed by the attention economy, as seen in the rise of clickbait headlines driven by PV-first mentalities.

One thing I think about mass media, especially terrestrial TV, is that they might be slightly underestimating the intellectual level of their viewers. When I talk about something slightly difficult, it's basically all cut. I'm not even allowed to use loanwords. In contrast, political and economic media on YouTube deliver difficult topics like causal inference exactly as I speak. Then, it resonates incredibly well with the audience that understands it.

I think terrestrial TV could estimate the intellectual level of viewers a bit higher. I think it would be fine to take a stance where viewers are encouraged to look up things they don't understand.

Personally, I completely agree that we could convey more specialized content, but in the process, the difficult parts tend to get cut first.

I've also been asked if I could rephrase things into other words so that older people can understand (laughs).

And when you do that, young people drift away (laughs).

That's why we need to reflect on whether the generational divide is really the fault of social media. Isn't it possible that newspapers and existing media are creating the factors that cause people to leave? I feel there might be a neglect of engagement with young people.

On the other hand, I think many viewers and users have become very adept at riding the "Behemoth." And since people who exploit it also emerge, the question is what to do about that. I feel that if we let the world become a place where such people win, it will become truly bleak.

That's true. I think the attention economy accelerates the worldview of "caveat emptor" (buyer beware). In short, it's the idea that if something happens after you buy a product, it's your responsibility for not being careful.

However, in the field of consumer protection, that is supposed to be considered wrong. Precisely because there is a large power imbalance and especially information asymmetry between the consumer who buys and the company that sells, consumer issues have aimed for resolution by requiring companies to do what they should as sellers.

Yet, when it comes to the space of public discourse, it's treated as the recipient's responsibility to improve their literacy and manage themselves. Even if the sender lines up lies, it's considered self-responsibility because the recipient chose to believe it. And they profit from that. But I think that's "unscrupulous," so the question is how to make it a slightly fairer space.

At Keio's X (Cross) Dignity Center, we are moving forward with a private-sector-led project to address challenges in the information space ecosystem. There is no silver bullet, but I believe platforms, media, advertisers, and we at the university must collaborate even further.

Thank you all for your time today.