Can the Creator Economy Survive in a Flood of AI? — A Deep Dive into MrBeast, Seedance, and the Value of the 'Authentic'

Introduction: February 2026, a Turning Point Indicated by Two News Stories

During one week in February 2026, I read two news stories and found myself lost in thought in front of my screen for quite a while.

The first was the news that the world's biggest YouTuber, MrBeast (Jimmy Donaldson), had acquired the fintech startup 'Step.' His channel has over 450 million subscribers, and a single video garners hundreds of millions of views. Why would he acquire a financial services company? The answer was inextricably linked to the shocking fact that his media business had recorded an estimated net loss of $80 million to $110 million in 2024 [2][3].

The second was the news that 'Seedance 2.0,' a video generation AI developed by TikTok's parent company ByteDance, had been accused of large-scale copyright infringement by the Hollywood film industry. The Motion Picture Association (MPA) condemned ByteDance with the words, 'Copyright infringement is a feature, not a bug' [10], and the actors' union SAG-AFTRA also strongly criticized it as 'unauthorized use of members' voices and likenesses' [9].

These two news stories appear unrelated on the surface. However, I could not help but feel that they were illuminating the same single, massive change from different angles. That change is the fundamental structural transformation of the creator economy.

As an AI venture executive and a visiting researcher at the Musashino University Asia AI Research Institute, I study the impact of AI on society every day. I am also involved in AI support for athletes as an advisor to the Shonan Bellmare Cycling Team, and I am at the forefront of new business development at the USEN WORK WELL AI Lab. From that perspective, what we are witnessing now is not merely a story of 'job automation by AI.' It is a civilizational turning point where the meaning and value of content itself are being questioned.

In this article, starting from the TechCrunch article 'Can the creator economy stay afloat in a flood of AI slop?' [1], I intend to thoroughly discuss the historical transition of the creator economy, the new threat of AI slop, the changing power dynamics between platforms and creators, and why the concept of 'Authenticity' will become the strongest weapon in the AI era, incorporating multifaceted research data and my own insights.

Chapter 1: The Birth and Evolution of the Creator Economy — From the Blog Era to a '$1 Trillion Industry'

1-1. The Beginning of the Revolution of 'Individuals Publishing'

When talking about the history of the creator economy, many people start with the rise of YouTube and Instagram in the 2010s. However, its origins are much older.

In 1999, the blog service 'Blogger' appeared. This service, which allowed anyone to easily publish information without specialized knowledge, popularized the concept of 'individuals publishing to the world' for the first time [15]. Until then, information dissemination was the exclusive domain of 'institutions' such as newspaper companies, television stations, and publishers. Blogger broke that monopoly and created a new category called 'personal media.'

The blog boom of the early 2000s was a symbol of the democratization of 'anyone can become a writer.' However, the means for those bloggers to earn revenue were limited, and it was only after the appearance of Google AdSense in 2003 that the concept of 'earning from a blog' began to feel realistic. Even so, only a very small number of people could make a living from blogging.

The arrival of YouTube in 2005 and Instagram in 2010 further accelerated this revolution. In particular, the introduction of YouTube's advertising monetization program in 2007 made the lifestyle of 'making a living by creating content,' which had previously been a pipe dream, a reality [15]. From this moment on, the creator economy began to establish itself as an economic sphere rather than just a world of hobbies.

Entering the 2010s, the spread of smartphones further accelerated the creator economy. Everyone began carrying high-quality cameras and video recording equipment in their pockets, and the barrier to entry for content production dropped dramatically. As Instagram evolved from photo sharing to video, and then to Stories and Reels, new professional categories like 'Instagrammer' and 'influencer' became established in society.

The arrival of TikTok (initially musical.ly) in 2016 brought a new dimension to the creator economy. The short-form video format further lowered production costs and led to an explosive increase in the creator population, centered on teenagers. TikTok's algorithm adopted a 'discovery-based' mechanism that spreads excellent content regardless of follower count, making it possible for unknown creators to gain millions of followers overnight.

This concept, proposed by Paul Saffo of Stanford University as the 'New Economy' [15], grew to an estimated market size of approximately $104.2 billion in a joint study by NeoReach and Influencer Marketing Hub in 2021 [15]. And there are predictions that it will reach $235 billion in 2025 and $1.487 trillion by 2034 [12].

1-2. The Evolution of Revenue Models — From Advertising to 'Community'

The revenue model of the creator economy has changed significantly throughout its history. The early mainstream was the advertising revenue share provided by platforms. YouTube's AdSense and blog-based Google AdSense were typical examples, a simple model where revenue was distributed according to the number of video views or page views.

However, from the mid-2010s, the limitations of this model began to appear. The platform's advertising cost per mille (CPM) declined year by year, and the revenue earned for the same number of views decreased. In 2017, an event called the 'YouTube Apocalypse' occurred. Major brands viewed the display of their ads before and after extreme content as a problem and stopped placing ads on YouTube all at once. As a result, the revenue of many creators plummeted overnight.

Creators who keenly felt this 'vulnerability of advertising revenue' began to search for new sources of income. Patreon, which appeared in 2013, revolutionized the creator economy as a platform where fans could directly support creators through monthly subscriptions. If you could receive $10 per month from '1,000 true fans,' that would be $120,000 in annual revenue—this simple calculation gave hope to many creators.

Current creator revenue streams are diversifying as follows.

| Revenue Source | Overview | Representative Platforms |
|---|---|---|
| Ad Revenue Sharing | Distribution based on views/impressions | YouTube, TikTok |
| Brand Deals/Sponsorships | Direct contracts with companies | All platforms |
| Memberships/Subscriptions | Monthly fees from fans | Patreon, YouTube Membership |
| Tips/Gifts | Support during live streams | YouTube SuperChat, TikTok Gift |
| Product Sales (Merch/D2C) | Direct sales of own-brand products | Shopify, BASE |
| Digital Content Sales | Sales of paid articles, videos, music | note, Gumroad |
| Online Salons | Operation of paid communities | Discord, Patreon |
| Live Events | Revenue from offline events | General |
| NFT/Digital Assets | Sales of ownership of digital works | OpenSea, Foundation |
| Courses/Educational Content | Paid learning content | Udemy, Teachable |

Of particular note is the reality that brand deals account for 68.8% of revenue[12]. Ad revenue sharing accounts for a mere 7.3%. In other words, for modern creators, revenue from platforms is just a 'bonus,' and the core of the business lies in relationships with brands.

This shift has fundamentally changed the 'nature of work' for creators. While creators of the past were 'content creators,' modern creators have taken on the role of 'marketing partners for brands.' While this brings business stability to creators, it also creates conflict between 'commercial pressure' and 'creative freedom.'

1-3. The Uniqueness of the Japanese Creator Economy

The Japanese creator economy shares commonalities with global trends while possessing its own unique characteristics.

According to a 2025 survey, the Japanese creator economy market size has reached 2.0894 trillion yen, showing rapid expansion with a compound annual growth rate (CAGR) of 15.5% since 2021 [14]. The potential market is estimated at approximately 14.5866 trillion yen, and further growth is expected. This growth is driven by sectors such as product/merchandise sales, advertising and marketing related to video posting, and skill sharing, which together account for about 70% of the total market.

A notable unique feature of Japan is the existence of **VTubers (Virtual YouTubers)**. The VTuber industry, led by ANYCOLOR (Nijisanji) and COVER (Hololive), has many fans both domestically and internationally through a unique content format of live streaming using 2D and 3D avatars. This can be called an innovative creator model originating from Japan, where the psychology of Japanese creators who want to avoid 'showing their faces' is fused with anime and game culture. The VTuber market has established a unique position globally and functions as an 'export' of the Japanese creator economy.

Also, the existence of a platform called note is unique to Japan. As a text-centric content sales platform, a wide range of creators, including writers, researchers, and business professionals, sell paid articles. Representative Kato of note has stated that 'it is becoming difficult to create value based solely on the amount of information' [14], emphasizing the importance of 'information quality' and 'writer personality' in the AI era.

Furthermore, in Japan, platforms for illustrators and doujin artists such as pixiv and BOOTH have formed their own ecosystems. pixiv has reached tens of millions of monthly active users, with many creators, both professional and amateur, publishing and selling their works. This coexistence of 'derivative work culture' and 'commercialization' forms the uniqueness of the Japanese creator economy.

However, the challenges faced by Japanese creators are also serious. Many creators are dependent on platforms and do not have established ownership over their own content or audiences. There is always a risk that a creator's activities will become unstable due to changes in platform terms, account suspensions, or algorithm changes.

This statement should strike a chord with Japanese creators as well. Dependence on platforms is a structural vulnerability of the creator economy, and that problem is becoming even more serious in the AI era.

In addition, the issue of defamation is a serious challenge faced by Japanese creators. Creator agencies such as UUUM, ANYCOLOR, and COVER have established internal teams specializing in responding to defamation and aggressive behavior toward creators and are continuously implementing countermeasures [14]. However, the structural problem of SNS anonymity remains deep-seated, and creating an environment where creators can work with peace of mind remains an urgent task.

Chapter 2: The End of the 'Ad Revenue Model' and New Business Models Shown by MrBeast

2-1. The Paradox of 'Deficit' Faced by the World's Largest YouTuber

MrBeast (Jimmy Donaldson) is a symbolic figure of the creator economy. With over 450 million subscribers and single videos garnering hundreds of millions of views, he has been looked up to by many creators as the 'pinnacle of success.'

However, 2024 financial data revealed a shocking fact. While his media business generated an estimated $250 million in revenue, it incurred a net loss of $80 million to $110 million [2][3].

Why does such a large deficit occur? The reason lies in the abnormally high cost of MrBeast's content production. His videos are produced by mobilizing large-scale sets costing tens of millions of dollars, prize money for participants gathered from all over the world, and a huge staff. It is not uncommon for hundreds of staff members to be involved in a single video, with production periods spanning several months. It is said that the production cost alone for the Amazon Prime reality show 'Beast Games' reached hundreds of millions of dollars.

This shows that the model of 'if you make high-quality content, you can recover costs through ad revenue' no longer works. We have entered an era where no matter how many views you get, ad revenue alone cannot cover production costs.

This phenomenon is not just a problem for MrBeast. Many major YouTubers are facing a double whammy of soaring production costs and declining advertising rates. YouTube ad revenue, once said to be '100,000 yen per million views,' has now dropped significantly, making it difficult to make a living solely from ad revenue unless you are at a massive scale.

2-2. The Counterintuitive Idea That 'Content Is Marketing'

So, why doesn't MrBeast go bankrupt? The answer lies in the product business he is developing.

The chocolate brand he launched, 'Feastables,' recorded $250 million in sales and over $20 million in profit in 2024 [2]. This is a surplus that more than compensates for the losses in his media business. Furthermore, the valuation of his holding company, 'Beast Industries,' has reached $5 billion [4].

This statement sharply highlights the fundamental shift in the creator economy. For MrBeast, YouTube videos are merely 'advertisements to sell products.' Videos that get hundreds of millions of views function as a marketing channel that is overwhelmingly more effective than television commercials.

And this acquisition of Step can also be understood in this context. By providing financial services to his 450 million followers, he is transforming from a mere content creator into the 'owner of a fintech company.' This is the cutting edge of the creator economy, where creators evolve from 'media companies' to 'product and service companies.'

This 'content equals marketing' model is also permeating Japan. The case where cooking researcher Ryuji (with over 6.4 million total followers) collaborated with Lawson to introduce a 'Kakuni-meshi' (braised pork rice) recipe, contributing to sales promotion and brand awareness [15], is a classic example. Creators are evolving from 'people who make content' to 'people who drive brands.'

2-3. A New Funding Model in the Creator Economy

MrBeast's case anticipates the changes in the entire creator economy. Looking at the market as a whole, M&A of creator economy-related companies in 2025 recorded a record high of 81 deals, a 17.4% increase from the previous year [13]. VCs and institutional investors are beginning to enter the creator economy in earnest.

Particularly noteworthy is the emergence of creator funds like Slow Ventures. This is a model where VCs provide operating capital in exchange for a portion of a creator's future earnings, a new form of investment that treats creators like startups. Creators can obtain operating capital without selling equity, while sharing a portion of their future revenue with investors.

Also, major platforms like Webtoon are significantly expanding their creator support programs. They are implementing initiatives such as the 'Creator Residency Program,' which invites selected creators to their Los Angeles headquarters for a one-week intensive development program, and cultural experience programs that invite select VIP creators to South Korea [14]. This indicates that platforms are positioning creators not as 'content producers,' but as the 'core of the platform.'

However, this change has both light and shadow. Creator-specialized attorney Frank Pol says the following:

In other words, while the creator economy is 'industrializing,' the reality is that the benefits are enjoyed by a handful of top creators, platforms, and investors, while the vast majority of creators remain in an unstable position. According to Goldman Sachs' estimates, only 0.1% of creators are able to make a living through their own channels [15].

This 'creator economy gap' may expand even further in the AI era. Top creators who can master AI tools will dramatically increase their productivity, and the gap between them and other creators will only continue to widen. The ideal of democratizing the creator economy and the reality of the widening gap will likely surface as an increasingly serious contradiction in the future.

Chapter 3: A New Threat Called 'AI Slop' — The Day the Internet Rots

3-1. What is AI Slop?

Are you familiar with the term 'AI Slop'?

This refers to the flood of low-quality content generated in bulk by AI. 'Slop' in English means 'mud,' 'sludge,' or 'swill,' and it describes the poor-quality AI-generated content overflowing on the internet as if it were a flood of mud [5].

The hallmark of AI slop lies in the combination of high volume and low quality. Using generative AI, one can mass-produce articles, images, and videos in seconds. However, much of this content lacks fact-checking, original perspectives, or insights, serving as nothing more than a rehash of existing material. As these flood search engines and social media feeds, truly valuable content gets buried—this is the essence of the AI slop problem.

The scale is unimaginable. A survey from April 2025 reported that 74.2% of newly created web pages contained some form of AI-generated text[6]. In other words, more than three-quarters of the results you see when searching on Google today may contain AI-generated content.

AI slop is causing serious problems in specific genres. In fields where accuracy is a matter of life and death, such as medical, financial, and legal information, there is an increasing number of cases where inaccurate AI-generated content appears at the top of search results. This can cause direct harm to users' health and assets.

Furthermore, AI slop also appears in the form of "SEO spam." There is an increase in malicious actors attempting to manipulate search engines with massive amounts of AI-generated content to earn advertising revenue, and since 2024, Google has been investing significant resources into eliminating such "spam sites."

3-2. How do users feel about AI content?

According to a survey conducted by CNET, 94% of social media users have seen AI-generated content, but only **11%** felt it was "useful" [7].

What do these numbers mean? People are seeing massive amounts of AI-generated content, yet most of them feel it is "useless." In short, AI slop is wasting users' time, eroding trust, and lowering the "quality of information" across the entire internet.

Even more serious is that it is becoming difficult for users to distinguish AI-generated content. Henry Ajder, founder of Latent Space Advisory, states the following:

"A coin toss"—meaning we can only distinguish AI-generated content with the same probability as flipping a coin. This implies that the reliability of information on the internet is fundamentally shaken.

In an official letter discussing the outlook for 2026, YouTube CEO Neal Mohan positioned AI slop as the most critical issue for 2026.

AI slop has become such a serious issue that the head of the world's largest video platform explicitly mentions it as a threat to their own platform. While YouTube is strengthening technical and policy responses to identify and remove AI-generated content, the scale and speed of these efforts are not keeping pace with the expansion of the problem.

3-3. "Model Collapse"—The vicious cycle of AI eating its own waste

The AI slop problem has an even more serious aspect. This is the phenomenon known as "Model Collapse" [6].

AI models are trained using massive amounts of text and images from the internet as training data. However, as AI slop proliferates, much of the content on the internet becomes AI-generated. Consequently, next-generation AI models end up incorporating that "AI-generated content" into their training data.

This is akin to AI eating its own "waste" and deteriorating. Compared to first-generation models trained on diverse and rich text written by humans, second- and third-generation models trained on data containing significant amounts of AI-generated content gradually become capable of producing only "average" and "impersonal" output.

To address this issue, the AI industry is pursuing various initiatives. Google has developed a digital watermarking technology called "SynthID" to enable the identification of AI-generated content [21]. Adobe co-founded the "Content Authenticity Initiative (CAI)" and is promoting international standardization to attach provenance information to content [21].

However, these technical measures do not provide a fundamental solution to the problem. An article in MIT Technology Review points out that "people remain emotionally swayed by content even when told it is fake" [21], suggesting that technical identification alone cannot restore social trust.

An article in Forbes JAPAN describes this issue as a "structural flaw that threatens the sustainability of the entire AI industry" [6]. This suggests that the AI slop problem is not merely a matter of content quality, but could potentially hinder the development of AI technology itself.

3-3-2. The 'Ghostwriter Problem' — An Era Where We Don't Know Who Wrote What

The AI slop problem has another serious dimension. This is the 'ghostwriter problem'.

In the past, a 'ghostwriter' referred to a professional who wrote books or articles on behalf of a famous person. Now, however, AI is taking on the role of the 'ghostwriter.' Suspicions are emerging everywhere that blog posts by prominent executives, social media posts by popular influencers, and papers by famous researchers are being partially generated by AI.

The situation is particularly serious in the world of academic papers. Since 2024, there have been successive reports of papers published in peer-reviewed academic journals containing text suspected of being AI-generated. There are increasing cases where the 'acknowledgments' section of a paper includes a note 'thanking ChatGPT for its help,' or where characteristic expression patterns generated by AI are detected.

In the academic world, 'who wrote it' is just as important as 'what is written.' A researcher's name indicates the credibility of the research and where the responsibility lies. If an AI wrote the paper, who bears that responsibility? The academic community has yet to provide a clear answer to this question.

A similar problem is occurring in the world of journalism. Some online media outlets publish AI-generated articles as if they were written by humans. Readers believe they are reading the product of a human journalist's reporting and thinking, but in reality, it is nothing more than text generated by AI. This is a deception of the reader and an act that undermines the fundamental trust in journalism.

To address this issue, some media outlets have adopted a policy of 'explicitly stating when AI is used.' The Associated Press (AP) has developed guidelines that require the clear labeling of any content that uses AI. However, media outlets that have adopted such policies are still in the minority, and standardization across the industry has not progressed.

In Japan as well, since the beginning of 2025, there has been an increase in cases where articles suspected of being AI-generated have spread on social media and were later corrected. What attitude should we take when reading 'articles that might have been written by AI'? This question is now being taken up in educational settings as a new challenge for media literacy.

3-4. AI Optimizes for the 'Average' — Where Differentiation Dies

An article published in Forbes JAPAN sharply points out the essential limitations of AI.

This observation has very important implications for creators. AI can certainly generate a large amount of 'pretty good' content. However, in a world overflowing with 'pretty good' content, there is no value in being 'pretty good.' Differentiation is the source of value, and AI kills that differentiation.

This problem is also becoming apparent in the world of brand marketing. The apparel brand True Classic experienced an issue where Meta's AI system replaced ads targeting men aged 30-45 with AI-generated images of older women [19]. As a result of the AI generating an 'averagely effective' ad, the brand's identity was compromised.

A British creative agency reported that its staff count decreased by 14% partly due to the influence of AI [19]. However, at the same time, according to a survey by Upwork, the demand for AI video generation and editing skills has increased by over 300% year-over-year [20], and the demand for talent capable of mastering AI is exploding.

In other words, a polarization is occurring between 'people whose jobs are taken by AI' and 'people who use AI to create jobs.' At this turning point, what is required of creators is not to have AI 'make' things, but to 'master' AI to amplify their own uniqueness.

Column: 7 Signs to Identify 'AI Slop'

Identifying AI slop is difficult, but there are several characteristic patterns. Below are 7 signs that you should suspect AI slop.

1. Excessively uniform writing style — There is no change in writing style throughout, and no emotional ups and downs can be felt. Human writing has variations in style, such as parts that are excited, parts that are cautious, and parts that incorporate humor.

2. Lack of specificity — There are many expressions like 'many experts point out' or 'according to research,' and few specific names, dates, or numbers. AI tends to prefer vague expressions to avoid the risk of 'creating' specific facts.

3. Lack of counterarguments — The text does not include counterarguments or critical perspectives regarding the claims. Human thinking is dialectical and includes a process of questioning one's own claims and considering counterarguments.

4. Lack of personal experience — It lacks specific first-person experiences such as "I experienced..." or "When I did...". Because AI does not "have" experiences, it is difficult for it to generate personal anecdotes.

5. Lack of emotional contradiction — Human writing often contains emotional contradictions, such as "While I think X, I also feel Y." Because AI tries to maintain logical consistency, it rarely exhibits such contradictions.

6. Lack of cultural and temporal context — It lacks the "vibe" unique to a specific era or culture. While AI can mimic cultural contexts included in its training data, it is difficult for it to capture the "vibe" of the current moment.

7. Overuse of "In summary" or "In conclusion" — AI tends to overuse these conjunctions when structuring text.

Of course, the presence of these signs does not necessarily mean the content is AI-generated, and conversely, it is possible for AI-generated content to lack these signs. However, being aware of these characteristics serves as one indicator when evaluating the reliability of information.

Chapter 4: The Collision of Seedance 2.0 and Hollywood — The Minefield of Copyright

4-1. The shock of the video where Brad Pitt and Tom Cruise "fight"

In February 2026, a video went viral on X. The video, which appeared to show Brad Pitt and Tom Cruise fighting, was viewed over 3.2 million times, and many users could not determine whether it was real or fake [9].

This was generated by "Seedance 2.0," a video generation AI developed by ByteDance. Seedance 2.0 is an AI model capable of generating high-quality videos from text and images, and its high performance shocked the industry. However, at the same time, its ability to make real people "appear" without permission raised serious ethical and legal issues.

Seedance 2.0 can generate more than just footage of celebrities. It can mimic a specific person's voice, facial expressions, and movements—all with high precision. This has brought the issue of "deepfakes" to a new dimension. While deepfakes used to require specialized skills and a great deal of time, tools like Seedance 2.0 allow anyone to generate high-quality deepfake videos in minutes.

4-2. The Motion Picture Association (MPA)'s condemnation: "Copyright infringement is a feature"

In response, the Hollywood film industry moved quickly. On February 20, 2026, the Motion Picture Association (MPA) sent a strong Cease & Desist letter to ByteDance [10]. In it, the MPA asserted the following:

Consider the weight of these words. "A feature, not a bug" is an accusation that the copyright infringement was intentionally designed. In other words, it is an argument that Seedance 2.0 intentionally ingests copyrighted content as training data on a massive scale to reproduce its "essence."

The actors' union SAG-AFTRA also harshly criticized it as "a blatant infringement that includes the unauthorized use of members' voices and likenesses" [9]. Actors are facing an unprecedented threat where their faces and voices are learned by AI without permission and used to make them "appear" in videos they never participated in.

This is a fundamental issue of "personality rights" and "portrait rights" that goes beyond mere copyright. Your face and voice being used in videos where you say things you never said, without your knowledge—this is a serious violation of individual dignity.

4-3. A storm of copyright lawsuits — The all-out war between the creative industry and AI

The Seedance 2.0 issue is just the tip of the iceberg. In the creative industry, which accounts for approximately 8% of U.S. GDP, lawsuits regarding the unauthorized use of generative AI for training are occurring frequently [17].

According to the Harvard Business Review (Japanese edition), court rulings on fair use are divided, making a prolonged legal battle inevitable [17]. One judge stated the following:

This statement fundamentally challenges the claim by AI companies that 'training is fair use.' The recognition that there is an essential difference between a human learning from a book and an AI 'extracting training data' from a book is spreading in judicial circles.

Another judge, however, has issued a ruling that partially accepts the AI companies' arguments [17], indicating that legal interpretations are not uniform. This legal uncertainty is a double threat to creators. On one hand, there is the risk that their content will be used to train AI without permission. On the other, there is the risk that content generated using AI might infringe on the copyrights of others. Creators are being forced to operate within a legal minefield.

In Japan, this issue is also serious. Although the 2018 amendment to the Japanese Copyright Act established provisions that allow for the use of copyrighted works for AI training purposes to a certain extent, legal experts remain divided on whether this permits unlimited training on copyrighted works by AI companies.

4-4. Lessons from OpenAI Sora — The 'Emptiness of a Humanless World'

Prior to ByteDance's Seedance 2.0, OpenAI released the video generation AI 'Sora' in 2024. While it garnered significant attention immediately after its release, subsequent developments were harsh.

According to data from January 2026, Sora's download count recorded a 45% decrease compared to the previous month, showing a sharp drop of approximately 25% from its peak [11]. While the rise of competitors is a factor, a more fundamental issue pointed out is that users have begun to feel a 'humanless emptiness' in AI-generated videos.

No matter how technically sophisticated it may be, AI-generated footage lacks human intent, emotion, and attachment to a story. Viewers intuitively sense that 'emptiness' and drift away. This is not a technical problem, but a problem concerning the essence of content.

This lesson applies to AI video generation tools in general. Technical sophistication is no substitute for the 'soul' of content. What viewers are looking for is not 'beautiful footage,' but 'footage where someone is trying to convey something.'

Chapter 5: The 'Invisible Rulers' of Platform Algorithms

5-1. An Era Where Algorithms Determine a Creator's Fate

In the creator economy, platform algorithms function as 'invisible rulers.' No matter how excellent the content you create, if it is not evaluated by the algorithm, it will reach no one. Conversely, content optimized for the algorithm is spread in massive quantities regardless of its quality.

This problem of 'algorithmic dependency' is a fundamental vulnerability of the creator economy. Every time a platform changes its algorithm, the revenue and influence of creators fluctuate significantly. From the massive decline in organic reach on Facebook in 2012, to the changes in Instagram's hashtag algorithm in 2019, to YouTube's focus on Shorts in 2022—each time, many creators have been affected.

As of 2026, how are the algorithms of each platform changing?

**Instagram's algorithm (2026)** emphasizes 'unification toward Views,' 'from saves to sends (shares),' and 'originality and two-way conversation' [16]. Instagram head Adam Mosseri emphasizes that 'the most important signal for increasing reach is the send' [16]. In other words, content that makes people want to 'share with friends' is valued more highly than 'likes.' This is a shift toward prioritizing the 'empathy' and 'conversational nature' of content.

Additionally, original content without watermarks, videos under 3 minutes, and good account status are recommended, while reposted content from other platforms tends to be rated lower [16]. This indicates that Instagram is attempting to build its own 'unique creator economy.'

**YouTube's algorithm (2026)** evaluates content by quantifying click-through rates, watch time, and post-view actions [16]. 'Audience retention' is particularly important, and videos watched to the end are judged as 'high quality.' Furthermore, it has been pointed out that AI-used videos may stop generating revenue if they are judged as 'mass-produced,' and the 'quality' and 'originality' of content are being demanded even more [16].

YouTube is moving toward mandating labels for AI-generated content as a countermeasure against AI slop, and restricting ad delivery to mass-generated content. For creators, this means that the strategy of 'using AI to generate massive amounts of content' will no longer work.

**TikTok's trend report 'TikTok Next 2026'** highlights three trends: 'Reali-Tea,' 'Curiosity Detours,' and 'Emotional ROI' [23]. Users are seeking 'unprocessed, raw stories and behind-the-scenes footage,' and content that 'makes one feel human warmth rather than crafted perfection' is valued.

As specific success stories on TikTok, Oreo's channel content increased its share by 12% in 2025, and Duracell increased its follower growth rate by 483% by finding an unexpected connection with K-Pop [23]. These cases demonstrate how powerful 'authenticity' and 'unexpectedness' can be.

5-2. X (formerly Twitter)'s algorithm shapes 'ideology'

In February 2026, a study published in the scientific journal Nature revealed a frightening aspect of social media algorithms. The study suggests that the X (formerly Twitter) algorithm may have the potential to push users to be more 'conservative-leaning' [24].

According to the study, posts from traditional media are displayed 58% less in algorithmic feeds, while posts from political activists are displayed 27% more [24]. This suggests that the algorithm may be favoring specific ideologies or information sources, potentially steering users' political views in a particular direction.

For creators, this is a serious problem. Creators cannot control who their content reaches or the context in which it is displayed. The 'invisible hand' of platform algorithms is secretly manipulating the relationship between creators and their audiences.

Furthermore, since Elon Musk's acquisition of X, content moderation policies have changed significantly, reducing the platform's reliability as a 'safe space' for creators. Many creators are considering migrating from X, leading to a dispersion toward alternative platforms like BlueSky and Threads.

5-3. Breaking Free from 'Platform Dependency'—Owning Your Audience

In this situation, forward-thinking creators are exploring ways to 'break free from platform dependency'.

A representative strategy for this is building an email distribution list (newsletter). Newsletter platforms, represented by Substack, allow creators to directly own their readers' email addresses. This model, which allows content to be delivered directly to readers without being swayed by platform algorithms, embodies the concept of 'owning your audience.' Substack saw a significant increase in paid subscribers in 2025, with many journalists and experts moving to Substack independently of media companies.

Also, building a community using Discord is effective. Discord is a platform where creators can operate their own servers and communicate directly with fans. It is less susceptible to the influence of platform algorithms and allows for the building of deep relationships with core fans.

Furthermore, the importance of a website with a custom domain is being re-recognized. While social media accounts are 'borrowed' from platforms, a website with a custom domain is your own 'property.' A custom site optimized for SEO brings stable traffic from search engines and reduces dependency on platforms.

Detavio Samuels (CEO of REVOLT and Offscript Worldwide) states the following:

As these words indicate, to build a sustainable creator economy, it is essential to build long-term 'relationships' rather than temporary 'buzz.' Platforms change. Algorithms change. However, the relationship with true fans does not change so easily.

Chapter 6: The Ultimate Weapon Called 'Authenticity'—Creator Survival Strategies in the AI Era

6-1. 'Trust is the Most Important Currency in the AI Era'

Henry Ajder, founder of Latent Space Advisory and leader of the Generative AI and Business program at the University of Cambridge, states the following:

This statement hits the core of the creator economy in the AI era. As AI slop floods the internet and fake news and deepfakes shake the foundations of society's 'truth,' the feeling that 'what this person says is trustworthy' has become rarer and more valuable than ever before.

According to a report by the Ministry of Internal Affairs and Communications, cases have been confirmed where fake videos of Prime Minister Kishida using generative AI were spread on social media, and cases where different disaster footage was spread as misinformation during the Noto Peninsula earthquake [21]. In January 2024, the World Economic Forum identified 'misinformation' as the most severe risk expected over the next two years [21]. Instances of information manipulation using generative AI are being confirmed around the world, such as the circulation of deepfake videos in the Indonesian presidential election and fake audio robocalls impersonating President Biden during the U.S. presidential primaries [21].

In such an environment, the value of content posted by 'real humans' based on 'real experiences' is paradoxically increasing. As AI optimizes for the 'average,' human 'individuality' and 'experience' become the source of differentiation.

6-2. Users are Seeking 'Authenticity'—The Psychology of UGC

From the perspective of consumer behavior, the demand for 'authenticity' is clearly demonstrated.

According to marketing research, consumers tend to view user-generated content (UGC) as 2.4 times more authentic than content created by brands [25]. Furthermore, there is research indicating that 86% of consumers view UGC as a reliable indicator of brand quality [25].

Why do people perceive UGC as 'authentic'? It is because behind UGC lie purely human motivations such as the 'desire for self-expression,' the 'desire for social connection,' and the 'desire for empathy' [25]. AI-generated content lacks these 'human motivations.' No matter how precisely it mimics human writing styles, the authenticity of the motive—'why this was written'—is conveyed to the reader.

In Japan, 'staff influencers'—sales staff belonging to companies who disseminate information as influencers—are attracting attention [14]. At companies like 3COINS, Zoff, and Shiseido, staff members post product introductions and usage tips on their company's social media accounts, gaining many fans [14]. This approach is perceived by consumers as more accessible and reliable than information disseminated by official corporate accounts, generating high engagement. This is a prime example of the value of content created by 'real humans'.

TikTok's 2026 Trend Report also supports this trend. The 'Reali-Tea' trend reflects the psychology of users who seek 'unprocessed, raw stories and behind-the-scenes content' [23]. In an era where 'human warmth' is sought after more than 'crafted perfection,' 'perfect but empty content' generated by AI cannot move users' hearts.

6-3. A Coexistence Model for AI and Human Creators—The Competitive Advantage of 'Sense'

So, how should creators use AI? I believe it is important to view AI not as a 'competitor' but as a 'partner'.

An article from the World Economic Forum states the following:

In a study on copywriting using ChatGPT, it was shown that collaboration between humans and AI taught with a copywriter's know-how is more likely to produce higher-performing copy than ChatGPT alone [18]. In other words, what maximizes AI's capabilities is human 'sense' and 'know-how'.

Forbes JAPAN describes the leadership required in the AI era using the word 'sense'.

The ability to judge not 'Can it be made?' but 'Should it be made?'. The courage to say 'no' to the 'statistically plausible options' presented by AI. This is the 'sense' required of creators in the AI era.

Apple in the iPod era did not choose the 'better specs' (more buttons and functions) that AI would likely have suggested, but instead chose the emotional promise of '1,000 songs in your pocket,' succeeding with a simple physical form, a click wheel, and a clean interface [19]. This demonstrates the importance of 'sense'—that even if AI can draft options, it cannot teach you which one 'feels right'.

Teng Liu, an economist at Upwork, also states the following:

The demand for AI skills is exploding, but it is a demand for humans who can 'master' AI. It is not a demand for humans who are 'used' by AI.

6-4. The Strongest Moat: 'Community'—Connections Beyond Platforms

In the survival strategy for creators in the AI era, what I consider most important is 'community building'.

A community is not just a collection of followers. It is a group of people who share a creator's worldview and values, connect with each other, and build something together. Such communities remain unshaken even when platform algorithms change or when AI slop overflows.

As an advisor to the Shonan Bellmare Cycling Team, I am personally involved in building a fan community. By connecting athletes directly with fans and providing personalized experiences using AI, we are striving to build deep relationships that go beyond mere 'support.' What I have learned from this experience is that the strength of a community lies not in its 'numbers,' but in its 'depth.'

1,000 passionate fans are worth more than 1 million indifferent followers. This is a modern reaffirmation of the '1,000 True Fans' theory proposed by Kevin Kelly, founding executive editor of Wired magazine, in 2008. In an era where AI can generate massive amounts of content, deep relationships with a 'small number of passionate fans' become the strongest competitive advantage for a creator.

Research results indicate that 'trust' is becoming a key trend in influencer marketing for 2026 [14]. There is a growing tendency for companies to prioritize 'engagement rate' and 'community quality' over follower count when selecting influencers. This shows that the value of 'narrow and deep influence' is rising above that of 'broad and shallow influence.'

Yongsoo Kim, Chief Strategy Officer at Webtoon, states the following:

As these words suggest, excellent platforms view creators not as 'content producers,' but as the 'core of the community.'

6-4-2. What happens when 'people with sense' use AI — Learning from real-world examples

The word 'sense' might sound abstract. However, looking at concrete examples makes its meaning clear.

Example 1: Midjourney × Professional Graphic Designer

Even when using the same Midjourney prompt, there is a clear difference between images generated by a professional graphic designer and those generated by a beginner. Professionals make judgments based on experience and aesthetic eye, such as 'what kind of image fits the purpose,' 'what kind of prompt will produce the intended result,' and 'which parts of the generated image to keep and which to modify.' AI is a 'tool,' and it is human 'sense' that masters that tool.

Example 2: ChatGPT × Experienced Copywriter

As mentioned earlier, collaborating with an AI that has been taught a copywriter's know-how produces higher-quality copy than ChatGPT alone [18]. This demonstrates that the judgment of 'what to write' (sense) is more important than the technique of 'how to write' (AI's capability).

Example 3: AI Music Generation × Music Producer

AI music generation tools like Suno and Udio can generate songs from text. However, when professional music producers use these tools, they create tracks that embody the style and emotion of a specific artist, rather than just 'AI-generated songs.' Producers make judgments such as 'what kind of song is needed,' 'which parts of the generated song are good and which need improvement,' and 'how to edit and arrange it to become a finished product.'

Example 4: AI Video Generation × Video Director

When generating videos using Runway or Sora, video directors make judgments such as 'what kind of scene is needed for the story,' 'what kind of visual expression conveys the emotion,' and 'how to edit the generated footage to achieve the intended effect.' AI can 'generate video,' but it cannot understand 'why that video is necessary.'

These examples show a division of labor where AI handles 'execution' and humans handle 'judgment.' And what determines the quality of 'judgment' is 'sense'—that is, the totality of experience, aesthetic eye, contextual understanding, and sense of purpose.

In the AI era, what creators should polish is their 'sense.' It is not the technique of 'making AI create,' but the ability to judge 'what to make AI create.' It is the courage to say 'this is not it' to the 'statistically correct options' presented by AI. This is the essential value of a human creator that AI cannot replace.

6-5. Concrete suggestions for Japanese creators

Based on the considerations so far, I would like to present concrete strategies that Japanese creators should adopt.

First, it is a shift from 'quantity of information' to 'depth of experience'. In an era where AI can generate vast amounts of information, the value of 'providing information' itself is declining. What holds value instead is the 'experience that only you can tell.' Stories of failure, struggles, and the process of growth—these are your unique content that AI cannot generate. Personally, I have accumulated diverse experiences, from technical college to graduate school, sports science research, entrepreneurship, the sauna business, and the AI business. I believe this 'intersection of diverse experiences' is the source of my content's uniqueness.

Second, it is breaking away from 'platform dependency'. Building 'your own channel' that is not swayed by platform algorithms, such as email distribution lists, proprietary communities (Discord, Slack), and your own website, is an urgent task. In particular, email lists function as a 'last resort.' No matter which platform disappears, if you have an email list, you can contact your readers directly.

Third, it is the acquisition of skills to 'master AI'. Instead of letting AI 'create' for you, you 'master' AI to amplify your own creativity. By leaving research, data analysis, multilingual expansion, and content mass production to AI, creators can focus on 'essential creation.' Personally, in writing this manuscript, I have utilized AI as a research partner while writing the insights and text myself.

Fourth, it is 'deep diving into niches'. Because AI optimizes for the 'average,' human creators have an overwhelming advantage in niche areas. Content deeply rooted in specific communities, cultures, or specialized fields cannot be replaced by AI. 'Sports x AI,' 'Sauna x Technology,' 'Careers of technical college graduates'—these are niches that I can speak about deeply precisely because it is me.

Fifth, it is the utilization of 'collaboration'. Collaborations with other creators, experts, and companies create 'chemical reactions' that AI cannot produce. Unpredictable value is born when different perspectives and expertise intersect. The reason I serve as an advisor to the Shonan Bellmare Cycling Team is also to create a unique intersection of 'sports x AI x marketing'.

6-6. The ultimate strategy of 'accumulating long-term trust'

When discussing creator strategies in the AI era, what I want to emphasize most is the 'accumulation of long-term trust'.

In the world of social media, 'going viral' is often spoken of as a metric of success. However, buzz is temporary. Today's buzz is forgotten tomorrow. In an era where AI can generate massive amounts of content, the value of 'buzz' is declining even further. Content generated by AI can also 'go viral' if it is sufficiently optimized.

On the other hand, 'trust' cannot be built overnight. Trust is only accumulated by consistently disseminating information with integrity over many years, keeping promises, admitting failures, and sincerely facing readers and viewers. This 'accumulation of trust' is the asset of a creator that cannot be replaced by AI.

One of the creators I respect is a researcher who has been disseminating scientific content for many years. He does not intentionally create 'viral' content; he simply continues to disseminate what he thinks is important, carefully and sincerely. That accumulation has created hundreds of thousands of 'true fans.' His content has a 'weight' that is clearly different from what AI generates. It is 'intellectual integrity' born from the accumulation of years of research and thought.

This 'accumulation of long-term trust' is also rational from a business perspective. Creators who have accumulated trust have their existing fans automatically spread the word every time they announce new content. Even without spending money on advertising, trust generates 'word of mouth.' This is a sustainable competitive advantage for human creators that cannot be replaced even if AI generates massive amounts of content.

Also, 'long-term trust' serves as a 'guarantee' when creators take on new business challenges. When MrBeast launched Feastables, his 450 million fans thought, 'If MrBeast made it, let's try it.' This is a typical example where the 'asset of trust' accumulated over many years generated 'initial customers' for a new business.

I myself am disseminating content with this 'accumulation of long-term trust' in mind. As an AI venture executive, as a researcher, and as someone active at the intersection of sports and AI, continuing to sincerely disseminate what only I can talk about is my strategy in the AI era.

Chapter 7: The Future of the Creator Economy—Outlook for 2034

7-1. The road to a $1.487 trillion market

Even amidst the chaotic situation, the market size of the creator economy continues to grow phenomenally. It is predicted to reach $235 billion in 2025, andreach $1.487 trillion by 2034 [12]. Within Japan as well, the market size is predicted toexceed 10 trillion yen by 2034 [15].

Several structural trends are driving this growth.

First is thespread of smartphones and the democratization of internet access. Smartphone penetration is progressing worldwide, and people who once only consumed content are entering as creators. Especially in emerging markets such as Southeast Asia, Africa, and South Asia, smartphones have become the first internet devices, and entry into the creator economy is surging, especially among the younger generation.

Second is thedemocratization of content production through generative AI. With the spread of AI tools, it has become possible to produce high-quality content without specialized skills. This is significantly lowering the barriers to entry into the creator economy. On the other hand, as the AI slop problem shows, a challenge has also arisen where an increase in 'quantity' does not necessarily mean an improvement in 'quality'.

Third is theestablishment of a 'direct support' culture. A culture where fans directly support creators, such as Patreon, YouTube Membership, and paid note articles, is becoming established. This serves as the foundation for a sustainable creator economy that does not depend on advertising revenue. In particular, Gen Z and Gen Alpha naturally accept the behavior of 'paying money directly to creators they want to support,' and this trend is expected to strengthen further in the future.

Fourth is the maturation of Web3, NFT, and blockchain technologies. After the initial bubble settled, NFT and blockchain technologies are becoming more practical as mechanisms for creators to clarify ownership of their digital works and earn revenue from secondary distribution.

7-2. The "Co-evolution" of AI and Creators — A New Form of Creativity

What I am most focused on is the phenomenon of the "co-evolution" of AI and creators.

AI is evolving as a tool to "expand" the creativity of creators. Musicians use AI to generate instrument parts they cannot play themselves. Writers use AI for brainstorming ideas and experimenting with writing styles. Video creators use AI to achieve visual expressions that were once impossible. Illustrators use AI to accelerate the conversion from rough sketches to finished products.

This multiplication of "human creativity × AI capability" is creating an entirely new form of creativity. It is a third category that is neither content generated by AI alone nor content created by humans alone.

The World Economic Forum states the following in an article titled "Why human creativity is becoming more important as AI evolves":

In this process of "co-evolution," what is required of creators is "AI literacy"—understanding the capabilities of AI and using it as an extension of their own creativity. This is a matter of adapting to new tools, just as photographers learned "how to use light" when the camera first appeared, or as musicians learned "digital music production" when DTM emerged.

Upwork data illustrates the concrete form of this "co-evolution." Demand for AI video generation and editing skills has increased by over 300% year-over-year, AI integration skills are up 178%, data annotation and labeling skills are up 154%, and demand for AI image generation and editing skills has recorded a 95% increase [20]. These figures show that the demand for "humans who can master AI" is exploding.

7-3. The "Economy of Trust" — What AI Cannot Create

Ultimately, I view the future of the creator economy as an "economy of trust."

In an era where AI can generate massive amounts of content, the value of "content itself" will decline. However, the answers to questions like "who made this," "why was it made," and "can I trust this person" cannot be provided by AI.

Henry Ajder states, "At the moment, spotting AI-generated content is basically a toss-up for you and everyone else" [21]. In other words, we are already living in a world where we "don't know what is real."

In this world, the value of "trusted people" and "trusted information sources" is higher than ever. Creators who have sincerely shared information over many years and built relationships of trust with their readers and viewers can survive even in the flood of AI slop through the "asset" of that trust.

Conversely, starting to build trust from this very moment is the most important investment for creators in the AI era. One sincere article written today, one honest video filmed today, one earnest conversation held today—the accumulation of these things forms an "asset of trust" that AI cannot replace.

7-4. From the "Attention Economy" to the "Ownership Economy" — The Next Evolution for Creators

Forbes JAPAN describes the next stage of the creator economy with the phrase "from the attention economy to the ownership economy" [22].

The "Attention Economy" is a model where creators earn advertising revenue by capturing people's "attention." However, this model creates dependence on platforms and strips creators of their "ownership."

The "Ownership Economy" is a model where creators "own" their content, audience, and business. Through direct monetization, independent communities, and their own products, creators can build sustainable businesses without relying on platforms.

MrBeast's Feastables is a classic example of this shift to the "ownership economy." He converted 450 million people's "attention" into the "ownership" of his own brand. And now, through the acquisition of Step, he is exploring a new form of "ownership" in financial services.

This transition from 'attention to ownership' will become even more critical in the AI era. In an age where AI can infinitely increase the 'quantity' of content, the value of 'attention' will decline. However, the value of 'ownership based on trust' will rise. Deep relationships with true fans, unique communities, and proprietary branded products—these are 'assets of ownership' that AI cannot replace.

Conclusion—Where Do You Stand in the Flood of AI?

Let us return to those two news stories from February 2026: MrBeast's acquisition of Step and the MPA's accusation against Seedance 2.0.

What MrBeast demonstrated is the truth that 'the value of content lies not in advertising revenue, but in trust and influence.' He is attempting to monetize the trust of 450 million people in a new form: financial services. Even if his media business is in the red, that 'asset of trust' is worth $5 billion.

What Seedance 2.0 demonstrated is the paradox that 'while AI can imitate human creations, it cannot imitate human trust.' No matter how precisely it mimics Brad Pitt, it does not generate 'trust in Brad Pitt.' When the Motion Picture Association condemned 'copyright infringement as a feature,' it was also a warning to the entire AI industry.

The creator economy is now at a major turning point. The advertising revenue model is becoming a thing of the past, AI slop is overflowing, and platform algorithms are seeking 'authenticity.' In this chaos, the difference between creators who survive and those who do not lies not in 'whether they can master AI,' but in 'whether they have built up trust.'

As an AI venture executive, a researcher, and a content creator myself, I continue to face this question. AI is not an 'enemy' that will steal my job, but a 'partner' that expands my creativity. However, what makes that partnership work is my own 'sense' and 'trust.'

Let me quote Henry Ajder once again.

Indeed, we cannot romanticize it. We are now facing a situation unprecedented in human history where we 'do not know what is real.' However, that is precisely why the value of 'being authentic' is higher than ever before.

In the flood of AI, where do you stand? Let us start building that answer from today.

References

[1] Can the creator economy stay afloat in a flood of AI slop? | TechCrunch (Feb 22, 2026)

[2] MrBeast's media business lost $80M — but his chocolate made $20M | Mashable (Mar 11, 2025)

[3] MrBeast Lost $110M Despite 450M Subscribers, Creator Economy Crumbles | TechBuzz.ai (Dec 8, 2025)

[4] MrBeast's $5 billion empire runs on generosity—but at a cost | Fortune (Sep 26, 2025)

[5] How to prevent the flood of low-brow, low-quality 'junk' content and AI-generated slop? | JBPress (Nov 30, 2025)

[6] Putting the brakes on the flood of AI-generated content: The industry's all-out war to prevent model collapse begins | Forbes JAPAN (Feb 10, 2026)

[7] AI Slop Is Destroying the Internet. These Are the People Fighting to Save It | CNET (Feb 23, 2026)

[8] A Letter from the YouTube CEO: Looking Ahead to 2026 | YouTube Blog (Jan 22, 2026)

[9] Seedance 2.0: ByteDance says it's strengthening safeguards on its AI video model after backlash | NBC News (Feb 16, 2026)

[10] MPA Sends Cease and Desist to ByteDance Over Seedance 2.0 | The Hollywood Reporter (Feb 20, 2026)

[11] Downloads of OpenAI's video AI app 'Sora' plummet as competition in generative AI intensifies | SBBit (Feb 1, 2026)

[12] Creator Economy Market Size, Share, Growth, and Forecast 2034 | Market.us (2024)

[13] Creator Economy M&A Hits Record High In 2025 | Forbes (Jan 15, 2026)

[14] Survey Results on the Domestic Creator Economy (2025) | Mitsubishi UFJ Research & Consulting (Dec 10, 2025)

[15] What is the Creator Economy? Introducing Business Models and Use Cases | Content Tokyo (Dec 13, 2024)

[16] [2026 Latest] Complete Guide to the Instagram Algorithm! Official Announcements and Professional Operational Strategies | comnico (Jan 30, 2026)

[17] How Should Generative AI Companies and the Creative Industry Coexist? Lessons Learned from Successive Copyright Infringement Lawsuits | Harvard Business Review (Japanese Edition) (Sep 16, 2025)

[18] Human Creativity Becomes Increasingly Necessary as AI Evolves | World Economic Forum (Jan 2, 2026)

[19] AI Reduces Creative Talent by 14%: The Leadership of 'Sense' That Companies Need Now | Forbes JAPAN (Feb 21, 2026)

[20] Demand for AI Skills Such as AI Video Generation/Editing and AI Integration Surges | ZDNET Japan (Feb 9, 2026)

[21] Determining the Authenticity of AI-Generated Content is a 'Toss-up': Experts Discuss How to Survive in an Era of Shaken Trust | Forbes JAPAN (Feb 9, 2026)

[22] From Attention Economy to Ownership Economy: Structural Challenges Faced by Creators and Solutions | Forbes JAPAN (Feb 3, 2026)

[23] TikTok Next 2026 Trend Report | TikTok for Business (2025)

[24] Research Suggests X's Algorithm May Be Pushing Users to Become More 'Conservative' | Forbes JAPAN (Feb 19, 2026)

[25] The Psychology Behind User-Generated Content | Taggbox (Dec 10, 2025)