News has surfaced that the government plans to promote a joint public-private investment of approximately 370 trillion yen across 17 strategic fields, including AI and semiconductors, by FY2040. AI, semiconductors, space, quantum, digital & cyber security, content, foodtech, disaster prevention, healthcare, ports, and telecommunications. Looking at these terms arrayed side-by-side, it might resemble a typical growth strategy. Yet, stepping back a little, this is more than just industrial policy. It is a preparation for AI to become the physical body of society.

At the same moment, Google DeepMind published a paper titled "From AGI to ASI." Within, AGI is treated not as a destination, but as a transitional state toward ASI. Scaling, algorithmic paradigm shifts, recursive self-improvement by AI, and large-scale multi-agent collectives. In other words, AI is not evolving simply as a convenient tool. The prospect of AI improving itself, AI systems operating together like organizations, and taking on multiple judgments and actions within society is beginning to be discussed as a very real subject of research.

Here, a slightly uncanny sensation arises. This feels similar to the dawn of the internet, yet it is completely different.

The Internet Connected Humans

In the era of the internet, humans chose to connect. Building a website. Sending an email. Writing on a bulletin board. Writing a blog. Joining social networks. Searching. Buying on e-commerce. Posting videos. Creating a profile. To participate was to register an account.

Undeniably, the internet transformed society. The speed of information flow accelerated. Media became personalized. The distance between companies and customers shrank. Advertising became targeted. The way we work, shop, and build relationships changed. Yet, there remained a distinct sense of "participating." You opened a site yourself. You registered yourself. You posted yourself. You searched yourself. You connected yourself.

The internet connected human to human, human to information, and human to service. At the center lay human agency. Humans stepped into the net.

AI Moves into the Human Domain

The era of AI is different. Rather than humans participating in AI, AI enters into human lives, work, institutions, and decisions.

Writing text. Generating images. Summarizing meetings. Selecting sales leads. Evaluating job candidates. Predicting customer reactions. Reading medical images. Detecting factory anomalies. Optimizing logistics. Supporting children's learning. Monitoring the elderly. Guiding administrative procedures. Parsing disaster prevention data.

Here, while it appears humans are using AI, human behavior itself is simultaneously being read by AI. What words do we choose? Where do we hesitate? Which proposals do we refuse? Which images do we react to? What facial expressions do we make? What hours of the day do we tire? Which institutions cause us discomfort? In what moments do we feel secure?

AI begins to handle not just information, but human attitudes and reactions. At this point, the meaning of participation shifts. Participation in the internet era meant creating an account. In the AI era, participation becomes the very act of existing as a human within society.

Participating Before Choosing to Do So

During the early days of the internet, there was still a somewhat idyllic atmosphere. "Have you built a homepage?" "Started blogging?" "Are you on Mixi?" "Do you have a Twitter account?" "Are you posting on YouTube?" Participation, to some extent, was a choice.

Of course, as time progressed, bypassing the internet became increasingly difficult. Yet even then, there was initially a sense of opening the door yourself.

AI is different. Companies deploy AI. Municipalities deploy AI. Schools deploy AI. Hospitals deploy AI. Recruitment systems deploy AI. Sales management deploys AI. Urban infrastructure deploys AI. Appliances, cars, and sensors connect to AI.

At that moment, even if an individual has not registered with any AI, they are placed within an AI-mediated environment. Before choosing whether or not to use AI, you may be classified, assisted, predicted, evaluated, and optimized by it.

This is a qualitative shift in participation. You do not enter a service of your own accord; society itself is becoming AI-driven. And simply by existing in that society, humans are already participating.

What a 370-Trillion-Yen Growth Strategy Signals

The government’s 370-trillion-yen growth strategy seems to bolster this shift at a national scale. It encompasses not just AI and semiconductors, but digital and cyber security, space, quantum, foodtech, healthcare, content, disaster prevention, ports, and telecommunications. These are not mere industrial classifications. They are the fabric of daily life.

Eating. Moving. Working. Learning. Receiving medical care. Preparing for disasters. Playing. Expressing. Buying. Living. Communicating.

AI enters these domains. In other words, AI ceases to reside only inside computers, nor does it end with smartphone apps. AI steps into factories, ports, hospitals, schools, local governments, agriculture, disaster prevention, homes, and content spaces. AI begins to acquire a physical presence.

A physical presence here does not mean robotics alone. Sensors, semiconductors, communication networks, data centers, municipal systems, business applications, generative AI, agents, XR, and game spaces, administrative paperwork—all of these become the body of AI. The state is attempting to give AI a body.

What the Roadmap from AGI to ASI Signals

Meanwhile, Google DeepMind’s roadmap from AGI to ASI shows the speed of AI's internal development. Until now, the evolution of AI has been something where humans research, develop, train models, and implement services. From now on, this changes slightly. AI assists in AI research. AI writes code. AI designs experiments. AI reads papers. AI generates hypotheses. AI collaborates as multiple agents. AI distributes tasks in a manner close to human organizations.

In other words, the very speed of AI's development is poised to be accelerated by AI itself.

The internet took time to spread. It required physical communication lines. It required terminals. It required browsers. It required search engines. It required e-commerce. It required social media. It required smartphones. There were sequential stages.

AI, however, is already built upon the internet. The cloud is there. Smartphones are there. Social media is there. Enterprise systems are there. Administrative systems are there. Payment systems exist. Location data exists. Generative data of images, audio, video, and text exists.

AI does not need to connect society from scratch. It begins converting humans on top of an already connected society. Hence, the velocity is entirely different.

The Internet Was an Infrastructure of Connection

The internet connected humans. It was a massive shift. Speaking with distant people. Having access to information. Individuals broadcasting. Small shops selling nationwide. Obscure talents being discovered. Communities of interest being born. Words flowing without passing through legacy media.

The internet was an infrastructure of connection. Yet, the connected humans remained human. The ones writing blogs. Posting photos. Shooting videos. Commenting. Buying. Selling. Viewing. Searching.

To be sure, platforms datafied human behavior. Advertising and recommendation algorithms altered human desires and attention. Yet, at the core of the internet lay connection.

At the core of AI lies conversion.

AI Will Become an Infrastructure of Conversion

AI summarizes human words. AI transcribes human voices. AI predicts human actions. AI classifies human tastes. AI assesses human skills. AI assists human decisions. AI restructures human jobs. AI externalizes human memory. AI treats human discomfort as anomalous data points.

AI does not simply connect humans as they are. It converts humans into processable forms.

Text becomes vectors. Voices become feature values. Facial expressions become emotional estimation. Purchases become predictive models. Resumes become scores. Relationships become networks. Hesitation becomes churn rates. Silence is handled as non-response.

At this point, certain aspects of human life become more convenient. Easier to find. Easier to recommend. Easier to assist. Easier to diagnose. Easier to manage. Easier to optimize. Yet, simultaneously, something is lost. The moment humans are converted into processable formats, they become a fraction less human.

This "fraction" will become a significant concern in the coming society.

The Dawn of Human Data Infrastructure

This is precisely where SHIRO & Co. is looking. What becomes necessary in the AI era is not just AI models. We need infrastructure that addresses how humans are read, converted, and rearranged by AI—what we might tentatively call "human data infrastructure."

However, this does not refer to a database of personal information. Collecting name, age, address, company name, title, purchase history, and location data is just an extension of traditional CRM and ad delivery.

Human data infrastructure must deal with more ambiguous layers. Trust. Discomfort. Attitude. Context. Rhythm of life. Relationships. Hesitation. The things unchosen. Unspoken reactions. The moment a person closes off. The conditions under which they feel secure. The things they inexplicably refuse.

As AI enters society, this layer becomes truly vital. Humans do not operate on logic alone. Even a correct proposal may be rejected. Even a convenient system can feel stifling. Even an efficient institution can carry an unsettling feeling. Even a perfectly optimized space may find humans unwilling to reside in it.

If AI enters society without this layer being designed, humans will gradually be converted. And they may cease to notice they have been converted at all.

Content as a Field of Participation

An intriguing aspect of this growth strategy is that content is included as a strategic Field. Games, anime, manga, music, and video. Previously, these were treated as entertainment or cultural industries. Yet, when AI, XR, generative technologies, virtual spaces, avatars, and sensory devices overlap, content ceases to be mere creative works. It becomes an environment in which humans participate.

Games act as playgrounds and become spaces of daily living. Virtual spaces become places to test alternative bodies. Avatars become another self. Generative AI becomes a mechanism to externalize the inner self. Music and images become environments to modulate emotion. Social media becomes a place where folklore arises.

Humans do not merely consume content. Within content, they alter themselves. Within content, they build relationships. Within content, they re-situate meaning. Here, too, the meaning of participation is shifting.

Viewing becomes participation. Reacting becomes participation. Generating becomes participation. Being recorded becomes participation. Being learned by a model becomes participation.

Participation in the internet era meant posting. Participation in the AI era means reacting, being converted, and being repurposed.

Does the Option to "Not Participate" Remain?

Here, a question arises. In the AI era, is it possible to choose not to participate?

In the internet era, there were still people who "did not use the net." They did not own a smartphone. They did not use social media. They did not write blogs. They did not use e-commerce. Though inconvenient, it remained possible for a time.

Yet, in the AI era, to what extent is non-participation possible? AI enters company business systems. AI enters recruitment. AI enters municipal counters. AI enters hospitals. AI enters financial screening. AI enters schools. AI enters transport and disaster prevention. AI enters urban surveillance and management.

In such a world, while you might choose not to use AI, choosing not to be processed by AI becomes exceedingly difficult. This is the crucial point. Using AI is fundamentally different from being processed by AI, and in the coming society, the latter will be far more significant.

Preserving What Cannot Be Fully Converted

What SHIRO & Co. attempts is not anti-AI. Rather, it assumes that AI will permeate society. AI will not stop. AI is convenient. AI transforms industries. AI assists lives. AI extends human capabilities. Yet, that alone is insufficient.

If AI converts humans, we must consider how to preserve what cannot be fully converted. Discomfort. Silence. Margins. Tactility. Coincidence. Slowness. Futility. Hesitation. Presence. Objects of daily life. Feelings that have yet to find words.

These are difficult for AI to process, yet they are vital to being human. The things that do not become data. The things that do not become scores. The things that are not optimized. The things that cannot be explained, yet undoubtedly exist.

What is truly necessary in the AI era is not to force these things into data, but to design a Meaning Layer so they do not fade away.

Not the Next Internet, But a Different Civilization

AI is sometimes called "the next internet." In terms of transforming society, the comparison holds. Yet, the structure is entirely different.

The internet connected humans. AI converts humans. The internet was a technology for humans to enter the information space. AI is a technology for the information space to enter human lives. The internet broadened participation. AI alters the meaning of participation itself.

In the internet era, humans participated by creating accounts. In the AI era, humans participate simply by existing. This difference is fundamental.

Thus, viewing AI merely as an IT trend risks misjudgment. This is not a matter of software. It is a shifting of how society is operated. It is a matter of how humans are processed. It is a matter of where meaning is preserved.

Who Designs the Meaning Layer?

The 370-trillion-yen growth strategy gives AI a physical body. The roadmap from AGI to ASI demonstrates the speed of AI's internal development. Overlaying these two reveals what lies ahead.

AI is turned into infrastructure by states. AI is operationalized by enterprises. AI is put on a path of self-improvement by research institutions. AI extracts human reactions through platforms. AI touches human desires and emotions in content spaces.

In that world, how do humans remain? This is a question not yet sufficiently addressed.

Policies speak of investment amounts. Enterprises speak of productivity. Research institutions speak of performance. Platforms speak of engagement times. Markets speak of growth rates. Yet, how are humans being converted? What is lost after conversion? Where do we preserve what cannot be fully converted? This inquiry is not yet at the center.

SHIRO & Co. is looking at that very layer. Trust OS. Decision Stack. Undecided Engine. Protocol Publishing. Meaning Layer. Human Data Infrastructure. These are not disparate terms. They represent the observations and language for handling trust, judgment, discomfort, meaning, and the tactile qualities of life in an era where AI enters society and converts humanity.

Humans are Connected, and Then Converted

The internet connected humans. Over more than two decades, that connection became the premise of society.

AI appeared on top of this connected society. That is why it is fast. The data is already there. The cloud is already there. Smartphones are already there. Social media is already there. Payments are already there. Enterprise systems are already there. Daily life is already digitalized.

AI reads, converts, and redistributes humans on this foundation. This speed is distinct from the diffusion rate of the internet. There may not be much time left for humans to slowly adapt.

Precisely because of this, what we need now is not just methods to use AI. We need to observe how humans are made to participate in an AI society. We need to record how humans are being converted into data. And we need to design a Meaning Layer for what cannot be fully converted.

The internet connected humans. AI converts humans.

That conversion has begun, and we are already participating.

© SHIRO & Co.

First published: 2026-06-25