The looming AI society and the transformation of major industries: Destruction and creation by AI

Japanese

Introduction

Among the high technologies of recent years, AI has been particularly impactful in reshaping society, and many are feeling the wave of social change caused by AI. We believe that the AI solutions we have seen so far are just the beginning. Structural reforms that will shake up lifestyles, economies, businesses, regulations and values will occur, and we are steadily approaching a future in which everything, even society itself, will be equipped with artificial intelligence. In order not to be left behind in such a future, it will be necessary to redefine your value, transforming and adapting in anticipation of the looming AI society. In this report, we examine how the evolution of AI will impact industry and society, touch on trends in major industries and consider what changes may occur in each.

Suggestions and a summary of the information derived in this report

• As the societal adoption of AI becomes more widespread, the high value-added zone in the AI industry value chain will shift from the conventional infrastructure layer (semiconductors, cloud, data centres, etc.) to areas closer to the user side. This specifically applies to the areas of ‘producing applications and AI agents’ and ‘Run the Business (RTB)’, which involves everything from design to implementation in order to integrate AI into various industries and bring about change.
• Physical AI has the potential to be so impactful that it could change existing industrial structures and create new industries. Furthermore, it is conceivable that once physical AI becomes widespread, it will be integrated into society as a whole through the ‘AI of Things (AIoT)’, rather than the ‘Internet of Things (IoT)’. As a result, the concept of a society that could be called an ‘AI of Society (AIoS)’ may arise, in which groups of AI scattered throughout all aspects of life are bundled together under integrated AI systems (‘social AI’) that govern specific domains and autonomously support society and industry.
• The age of AI will be a period of significant social change. It will be important to identify the value that your company will create within the AI industry structure. It will then be necessary to determine which areas to focus on to create a competitive advantage through ‘architectural thinking’, which involves deciphering the entire industry structure and calculating backwards from user value.

Chapter 1: Industrial transformation driven by AI

1. New social redefinitions brought about by AI

The place of AI in a changing society

Since the late 2010s¹, the most important agendas for society, nations and businesses have increasingly included perspectives such as the environment, poverty, human rights, technological innovation, regional revitalisation and diversity. Through the pursuit of sustainability resulting from deterioration of the global environment, shifting values due to demographic changes and geopolitical divisions that create conflicts, the nature of the desired form of society is changing at a dizzying pace. It is believed that, within a few years, a future will arrive that the current social structure cannot accommodate.

Taking population decline as an example, it is estimated that Japan’s population will decline to approximately 70m, causing the domestic market to disappear, serious labour shortages, infrastructure/land vacancies and administrative collapse. The current social system will become dysfunctional, and a future in which the nation is in danger of decline cannot be brushed aside as an impossible scenario.

In order to adapt to a society with a changing structure, it will also be important to harness the full potential of AI, which is expected to augment, expand and substitute for human capabilities.

This report delves into the impact that AI will have on society and the future vision of industry, looking ahead to the transformation of the social structure expected over the next 10 to 20 years.

The impact of AI

Looking at recent AI trends, general-purpose and industry-specific AI are being created based on large language models (LLMs). With an eye towards incorporating AI into robots and various devices, a historically massive investment race is underway, focused on the foundation of AI services—semiconductors, data centres, power supplies, etc.

It is believed that as AI implementation rapidly progresses, existing value chains in the future will be disrupted and redesigned. Specifically, as routine intermediate processes are simplified or eliminated through automation by AI and other means, traditional multi-structured industries and systems that rely on manpower will gradually be eliminated, and a shift to new business processes based on AI implementation will be required. As discussed below, the main battleground for competition in the value chain is also likely to change significantly, and the key agendas of industries are also expected to undergo transformation.

Figure 1: The impact of AI

AI will be implemented at the core of society and industry, changing the way people and companies think and interact

Changes in key agendas across various industries

• Changes in business-to-business dynamics
  In addition to increasing the number of users, which is the traditional source of economies of scale, companies with vast computational resources and data can continuously train higher-performance models. This process satisfies the virtuous circle requirement of ‘number of users × data × computational resources × AI models’. It is believed that such companies will emerge as the core of a new scale economy based on AI, resulting in a surge in demand and massive investment centred on semiconductor players. Furthermore, with the rise of custom ASICs (application-specific integrated circuits) with in-house specifications, competitive and partner relationships within industries are likely to change dramatically.
• AI regulatory and safety requirements
  In order to incorporate advanced AI into social infrastructure and important operations, safety assessments with a series of governance systems will need to be incorporated. These must cover everything from the procurement and management of training data to the assignment of metadata for products and accountability to users. Indeed, the following developments are beginning to emerge.
  • The European Union (EU) has developed the EU AI Act, a framework that classifies AI systems according to a risk-based approach and imposes strict obligations on high-risk applications.
  • In the US, executive orders have provided guidelines for safety testing, model evaluation and the application of AI on critical infrastructure.
  • Regarding copyrights and displaying the source of data and materials used in generative AI, C2PA is also attracting attention as a technology for ensuring the authenticity of generated content and transparency of the generation process.
• Economic changes and constraints regarding AI costs
  Training large models requires enormous amounts of GPU processing time and power, and training costs are on the rise. On the other hand, although the inference cost of trained models is on a downward trend due to performance improvements in hardware and software, the total cost is on an upward trend due to facility expansion in response to the sharp increase in demand. As a result, the supply shortages of power and GPUs/memory are becoming a constraint on market growth.

Changes that will occur in societies and industries reshaped by AI

In a society where AI continues to permeate the world and everything becomes equipped with AI, the way we perceive human values and the nature of society itself may change.

The more AI substitution advances, the more people will be forced to redefine specifically what it means to be human. This is because it is believed that the rational solutions that AI provides are not necessarily the optimal solutions for humans, and that irrational decisions that respect human values, empathy, emotions and ethics will become more important than ever in making such decisions. This is why the humanistic perspectives of ‘can people live like human beings?’ and ‘is human dignity being compromised?’ will become prerequisites for the selection, operation and institutional design of AI technologies.

Thus, the more AI permeates industry and society, the more it will transform those existing ways of being.

Disruption and creation of traditional value ecosystems

As AI continues to permeate throughout society and amidst the changes that may occur along with the restructuring of existing value chains in industries, it is necessary to pay attention to how the centre of gravity (source of profit) of the AI industry’s profit pool will shift.

Figure 2: Shifts in the main battlegrounds of the value ecosystem

Currently, the major battlegrounds include model development, semiconductors and data centres, which are essential for providing AI services. Furthermore, massive investments are concentrated in technological competition for AI performance and infrastructure enhancement to acquire computing resources. However, once these technologies and supply capabilities reach a certain level and the market enters a phase of progressive generalisation, price competition will intensify. At the same time, there will be a risk of oligopolisation, making it difficult to secure high profits from the infrastructure layer alone. Therefore, in order to survive in the AI market in the medium to long term, it will be important to be able to discern changes in winning strategies that extend beyond the overheated investment battle.

As such, the next key areas to focus on are ‘producing applications and AI agents’ and ‘Run the Business (RTB)’, which involves everything from design to implementation in order to integrate AI into various industries and bring about change. With massive investments currently underway, some fear that the AI bubble could burst if the value created by the AI infrastructure built through these investments fails to translate into monetisation models or use cases sufficient to recoup the costs. We therefore believe that for the AI market to thrive in the future, it will be important to have players who can handle the ‘Run the Business’ aspect—those who can consistently design, implement and operate services, functions and systems that deliver the values demanded by society and users.

Although there are currently no key players in this role, we are beginning to see signs of a battle for supremacy, such as with the use of physical AI with robots. In this area, we see a value proposition that will lead the future AI market, and together with our clients, we will take a rapid approach to creating a new era.

2. From generative AI to physical AI and beyond

The spread and evolution of AI

Although interactive generative AI has only recently become widespread worldwide, there are now growing expectations that agent AI can autonomously cooperate with multiple AIs, perform human-like responses and judgements, and act as a partner to humans in a variety of daily life and industrial situations.

Furthermore, the evolution of AI is not limited to the digital space, but is also expanding into the physical space in conjunction with robotics and IoT. Physical AI refers to technology that recognises the physical laws of the real world, acting and making decisions autonomously in physical space. AI is moving beyond information processing in virtual space and is entering a phase where it interacts with the real world through entities such as self-driving cars and robots.

Changes in the structure of industry due to physical AI

By adding a recognition of the physical environment brought about by physical AI to the digital and autonomous data processing and decision-making of agent AI, it will be able to truly ‘act in reality’ while analysing and extracting the physical impact of the position, weight, condition, etc., of objects. In addition, its use in fields such as manufacturing, mobility and medical care is expected to accelerate.

Figure 3: The impact of physical AI

• Manufacturing
  Traditionally, industrial robots have been used on production lines, and some robots utilise AI to make rule-based decisions. Unlike those robots, physical AI can perform tasks based on autonomous decisions. For example, it is expected that a humanoid robot incorporating physical AI will be able to recognise the size, shape, material and welding condition of an object while welding in the manufacturing process. It will then autonomously determine how much pressure and time to apply to which part, without any manuals. In addition, it may be possible to accurately reproduce skilled techniques and craftsmanship, which were considered tacit knowledge, by incorporating them into data.
  We can also envisage a future in which factory knowledge, obtained through repeated simulations using AI digital twins, is defined in software and applied to each factory as a ‘manufacturing OS’. In the future, we can expect the formation of a new industry using these manufacturing OS as a medium.
• Mobility
  When physical AI is introduced to vehicles or aircraft, it will be able to autonomously judge and execute route changes and vehicle movements according to the situation based on pedestrian and obstacle information captured by sensors and the surrounding traffic conditions. In the field of logistics, non-stop delivery is being considered in which warehouses, vehicles, drones, etc., dock through interlocking, and dramatic efficiency improvements are expected.
  Asynchronous connection of physical and spatial information, such as local traffic signals, cameras, sensors and buildings, will make it possible to detect situational changes in real time, such as detecting pedestrians in blind spots or behind buildings, as well as sudden accidents or incidents. If this happens, mobility in society may take on a new role as social infrastructure, not just for the transportation of people, but also for monitoring, disaster prevention and the maintenance of society.
• Medical care
  Examples of possible uses of physical AI in medical care include robots autonomously learning advanced procedures such as suturing with ultra-fine threads and manipulating intravascular catheters, or learning surgical information such as facial expressions and trauma through repeated simulations. As a result, it is conceivable that robots could take over medical procedures traditionally performed by humans.
  The scope of use of physical AI is not limited to treatment. For example, by stepping into people’s daily lives and acquiring and analysing biological information (body temperature, blood pressure, brain waves, etc.) and activity information (walking, falls, meals, etc.), it is possible to propose diet and exercise menus that take into account medium- to long-term health, and to support people’s lives by providing physical assistance. In this way, we may see a shift towards a ‘lifestyle intervention industry’ in which robots play a role in preventive medicine.

As a result of the utilisation and application of AI, it is thought that an ‘AI of Things (AIoT)’, rather than an ‘Internet of Things (IoT)’, will spread, and that AI will become an integral part of our lives, directly solving real-world problems as a substitute for humans. As an example of this future, we can envision the arrival of an age in which AI systems scattered across various domains are integrated into unified AI systems (‘social AI’) that govern specific domains and autonomously support society and industry. As such, if integrated AI becomes responsible for certain areas of social functionality, it may not be long before the concept of an ‘AI of Society (AIoS)’, in which AI is implemented at the core of society, becomes commonplace.

Figure 4: The trend from generative AI to physical AI

3. Benefits and threats of AI

Benefits and threats of AI

There is no doubt that AI will bring many benefits to human society in the future.

In addition to providing high added value to both routine and non-routine work, including creative work, it is expected to significantly improve the productivity of intellectual labour and autonomously handle everything from task instructions to completion reports. This could be one of the measures that help solve labour shortages, which is a major issue, especially in developed countries.

Additionally, AI has the advantage of economies of scale. In other words, the more we use it, the more data is accumulated, the more performance improves and the more inference cost decreases. The more we benefit from AI, the more opportunities and scope will expand at an accelerating pace.

However, there are also things to look out for as AI advances. For example, if input data is biased, issues such as hallucinations (answers that seem to be correct but are wrong), unethical answers and the decision-making process becoming a black box can frequently appear, and output without human intervention can be a serious risk. At the corporate level, cases have already begun to emerge where companies have disclosed output without noticing hallucinations. As AI further permeates throughout industry and society in the future, these impacts are likely to become even more serious.

We also need to look at changing infrastructure constraints and oligopolistic market risks. GPUs/NPUs, data centres, telecommunications and power are important components of the key infrastructure systems for AI, and competition for them is intensifying. As a result, the market is becoming a place where success or failure depends on how well a company can procure and capitalise on these strategic materials. At present, there are only a limited number of players with these capabilities, and a trend towards an oligopolistic market centred on big tech companies has already begun to emerge. This may lead to issues such as higher costs for downstream players and geopolitical risks.

Figure 5: Benefits and threats of AI

4. Understanding the future of the AI industry

The disappearance of industry frameworks and the industrial architecture that should be recognised

As the environment surrounding society and industry changes dramatically due to the catalyst that is AI, we believe that in order to find new value, ‘architectural thinking’, which involves understanding the entire industrial structure and working backwards from user value, will become increasingly necessary.

In the AI era, which will be a period of great social change, it will become extremely important to identify the value that your company can create within the AI industry structure and determine which areas to focus on to create a competitive advantage.

Figure 6: AI industry architecture from a functional perspective

The way value itself is perceived will also be considered important. The PwC Japan Group categorises the AI industry structure into three layers: social value, industrial value and functional value.

Figure 7: AI industry architecture formed by value points

^{* New value domains proposed in ‘Value in motion’, a research report that investigates and analyses how society and industry will change as a result of AI technological innovation, and what kinds of value companies will create within that context}

Values designed with architectural thinking will transcend industry boundaries. Therefore, instead of the traditional ‘cross-industry’ approach of cooperation between companies to realise specific themes, we believe that the mainstream approach for the future growth of the AI industry will be the expansion of ‘value solidarity’. In this approach, the layers of value to be implemented are defined and the key players and assets necessary for this purpose are gathered.

Redefining the management agenda required in the AI era

With the transformation of traditional value chains due to AI, the management agendas to be addressed by top management will change drastically. It will be important to address the following points in particular.

Figure 8: Redefining the management agenda in the age of AI

❶ Competition and revenue strategies: Rethinking competitive markets and cash points

It is important to identify the areas of value creation worth pursuing within the AI value chain and redefine competitive advantages based on your company’s core competencies. It is also necessary to update your pricing and revenue structure to adapt to the ever-changing competitive environment.

❷ Business strategies based on constraints: Develop tactics based on AI constraints and the AI ecosystem

In order to survive in an increasingly competitive environment, it is essential to identify the infrastructure constraints related to your own company, such as data centres, GPUs or power supplies, and then develop strategies for securing victory in the medium to long term. Given these constraints, partnering (M&A, alliances, etc.) is an essential approach to securing a competitive advantage in the massive AI market.

❸ Global standard architecture: Establishing a multi-regional AI architecture

In order to achieve global implementation in the AI era and reliably exceed the break-even point for AI solutions, where the initial investments are likely to be huge, AI solutions will need to come standard with mechanisms such as multi-regional RAG (Retrieval-Augmented Generation). This approach distributes systems across multiple regions, to ensure low latency, high availability and fault tolerance. There will also be a need to develop ‘Agent of Agents’, AI agents that evaluate and check AI applications and AI agents.

❹ User interface creation: Mass production of monetisation use cases

In the AI market, which is built on massive investment in semiconductors, etc., the top priority will be to create a highly profitable monetisation model that will allow a company to recoup their investment. As evidenced by recent talks of Japanese companies looking to acquire robotics companies, the extent to which AI players themselves can design user interfaces that lead to value creation will be an important factor in determining the success or failure of the market.

❺ Proprietary data strategies: Establishing proprietary data to support competitive advantages

In the AI era, relying solely on general-purpose or open data that can be used by other companies will not be enough to secure a sustainable competitive advantage. It will be extremely important to be able to accumulate and utilise data assets that only your company has access to. The key will be how to develop a proprietary data cycle that includes securing large amounts of unique data leading to the creation of high added value in specific industries or businesses, ensuring reliability, incorporating it into AI within your own services and creating value.

Chapter 2: Changes coming to various industries

So far, we’ve discussed the evolution of AI, its benefits and threats, and its impact on society. From here, we’ll delve into exactly how the expansion of AI can impact industries around the world.

1. Telecommunications industry

Infrastructure constraints become bottlenecks in an AI society

First, let’s consider the impact of AI proliferation on telecommunications infrastructure.

It is expected that the spread of AI will require the construction of telecommunications infrastructure that can process traffic data, which will increase dramatically, at high speed and with large capacity, ultra-low latency and low power consumption. In fact, it is estimated that network traffic will approximately double² between 2025 and 2031 due to the fact that video content can easily be mass-produced via generative AI and that real-time data can easily be transmitted and received.

In order to meet these demands, it will be necessary to establish a system that enables simultaneous connection of a large number of devices with ultra-low latency, autonomous network switching across a wide bandwidth and high-speed communication. However, given the considerable amount of time required for telecommunication companies to improve the quality of communications, which has become a major issue in recent years, this is not something that can be achieved overnight and will be a significant burden for companies.

In addition, as worldwide demand for computational processing is expected to grow rapidly (some estimates suggest an approximate 70-fold increase from 2018 to 2030³), ‘edge AI’, which allows for the distribution of computational processing, is attracting attention. Edge AI refers to technology that performs AI processing on devices close to the user rather than in the cloud. An example of edge AI is the integration of AI in user-side personal computers and smartphones. In the telecommunications industry, there are efforts to add computational processing functions to communication base stations. With this, computations through data centres are distributed and reduced, as processing is carried out at the communication base stations. Since large amounts of power are required for computations, this results in power consumption being distributed, leading to benefits such as reduced disaster risks and loads on power transmission and distribution facilities. In addition, from the user’s point of view, there is the advantage of low-latency processing, as data does not need to travel back and forth to large DCs located far away. Because of its low-latency characteristics, edge AI truly shines in use cases that require real-time performance, such as autonomous driving and physical AI. Furthermore, telecommunications carriers will become increasingly important players responsible for building the ‘communication facilities × edge AI’ infrastructure that will support such use cases.

As such, the telecommunications industry has an important role to play in building the infrastructure for an AI-driven society. Failing to build this infrastructure could create a bottleneck in AI adoption, which will support critical operations in the telecommunications industry, and ultimately risk undermining corporate competitiveness and market growth.

2. Manufacturing industry

The future of factories lies in software

Next, we will look at the impact of AI on the future of the manufacturing industry.

The benefits of AI in manufacturing factories are expected to be seen, for example, in autonomous production and dramatic improvements in supply and demand forecasting. If AI and robots can autonomously operate factories—maintaining production lines with low defect rates and creating and executing lean plans based on the identification of best-selling and underperforming products, etc.—without the need for human operation or judgement, operations with extremely reduced inventories can be expected.

With the implementation of AI into the manufacturing process, the data and know-how accumulated by manufacturers, including the refined skills and craftsmanship passed down to only some craftsmen, can be simulated using digital twins and maintained as templates, which could lead to entire factories becoming software-driven (a manufacturing OS). ‘Manufacturing OS’ here refers to an AI-powered digital platform that defines manufacturing technologies and processes that tend to be tacit knowledge of companies, conditions for minimising defective products and methods for adjusting production plans according to fluctuations in supply and demand, all at the software level. This enables the same level of quality at any factory.

Through a manufacturing OS, even tacit knowledge, such as the skills of craftsmen scattered across various factories, may be converted into knowledge (assets) that can be reproduced across multiple factories within the same company. As a result, even small businesses and other companies that previously struggled to establish production systems may be able to acquire superior manufacturing capabilities, technologies and networks by using a manufacturing OS.

In addition, the traditional business model of mass production at major manufacturing sites and delivery to each region while holding inventory can be shifted. This shift is driven by improved supply and demand forecasting using AI, enabling a business model focused on order-based production without the need to keep inventory. In the future, small-scale manufacturing bases may be created in each region where orders are received, forming a satellite manufacturing network.

These developments are expected to have a ripple effect on the logistics industry. The formation of this new manufacturing network could lead to a sharp increase in the need for last-mile delivery of products made to order at satellite factories scattered around the country, which could lead to the restructuring of regional distribution networks. In the past, the competitiveness of logistics companies depended on factors such as the number of personnel and trucks, but in the future, this competitiveness may be determined by the ability of AI to build logistics systems that minimise waste.

In this way, the spread of AI in the manufacturing industry could impact even the logistics value chain along with supply chain transformation.

3. Entertainment industry

AI will change the future of entertainment production

Finally, let’s look at the impact on the entertainment world, which is deeply connected to the lives of ordinary consumers and users. Reflecting on the relationship between entertainment and our daily lives, much of the time we spend watching TV and movies has shifted to watching them on our smartphones, and watching other videos on smartphones, including on social media, has already become the norm. What kind of changes are taking place in the areas of production and distribution behind this trend as AI becomes more prevalent?

A quiet seismic shift has already begun to take place in entertainment production. Advances in AI have led to efforts to semi-automate script, video and music production, as well as to create videos from existing data without the need for on-location filming. The production of scripts, videos and music, which used to require lengthy processes and budgets of several hundred million yen, can now be completed in a shorter time and at lower cost thanks to the semi-automation of generation and editing.

This movement is expected to steadily continue and become a clear trend within a few years. Due to lowered barriers to production, the number of people entering the entertainment industry will explode, and we can expect an era of an ‘oversupply’ of content. The quality of production sites will be determined by how well AI is utilised. It is expected that players who can appropriately manage which AI stacks to adopt, how to train them and how to operate them will generate high added value.

Despite the positive impact of lower production costs, this change could mark the beginning of an undesirable future for established production companies. Traditionally, production sites have received work from TV stations and distribution platforms, using production costs as their source of funding and profits. However, if production costs are reduced due to efficiency improvements made by AI, the unit price for contracts will drop, squeezing profits. Platforms themselves may also use AI to internalise production, leading to a potential future where production companies face a lack of work.

The key to survival lies in the use of intellectual property (IP) and how well AI can be handled. Specifically, the following four points will be pivotal.

1. Developing IP strategy in-house: Can you diversify revenue streams through merchandise, events, game adaptations and overseas expansion while maintaining ownership of works by embracing planning, rights and finance?
2. Building an AI stack: Optimise the generation model + evaluation model + workflow automation according to your company’s specifications to create an AI blueprint (pipeline) from planning to final output. Can you select models, fine-tune, systematise prompts and log quality standards to provide a foundation for producing high-quality products while taking IP/rights into consideration?
3. Managing fan communities: Design fictional universes and participation experiences to ensure recurring revenue after production. Can you create value not only from the video itself but from the experience as a whole?
4. Keeping existing IP/copyrights in mind: It is necessary to bear in mind that AI-powered content creation may infringe on existing copyrights and the rights of rights holders (scripts, actors, voice actors, etc.). Can you put a system in place that can properly address this in production?

The question is whether production companies will be able to break away from being mere subcontractors and take control of the rights to works themselves, generating revenue through secondary use (merchandise, events, game adaptations, overseas expansion, etc.). This will require the production companies themselves to perform some of the functions that they traditionally depended on the TV stations for: financing, management of rights and sales strategies. With the advancement of AI, the question is whether production companies in the entertainment industry will be able to evolve into creative companies that ‘use AI to increase efficiency’, ‘own IP and connect with fans’ and ‘incorporate planning, rights and finance’.

AI will change the future of entertainment distribution and advertising

Let’s next look at what changes are taking place regarding the distribution of video.

Changes in viewing trends and new trends in distribution: An age ruled by AI and algorithms

Consumer viewing habits are changing dramatically. In the past, it was common to watch TV in a fixed location, but now the on-demand format, which allows people to enjoy their favourite programmes whenever and wherever they want, is the norm. This change has affected the distribution structure of video content, and the traditional mass media-centred model is losing its relative prominence.

Recently, social media and video distribution services have become the main places for watching content, making searching increasingly unnecessary. There is now a system where AI analyses preferences and emotions using ‘invisible algorithms’ and then delivers content automatically. The integration of viewing and purchasing is accelerating, and the nature of advertising is about to change on a fundamental level.

The position of advertising companies and existing models

The value of ad space itself remains. In fact, for social media and video streaming services, a platform’s advanced targeting enables advertisers to reach directly to those who are most likely to understand their products, and ad space remains an important resource for meaningful advertising.

The concern, however, lies with the existing mass media and advertising agency model. The current system of inserting the same advertisements between programmes for a mass audience will continue to play an important role in raising awareness and improving brand image. However, this is not competitively advantageous in driving conversions and sales, which lead to product purchasing behaviour, compared to hyper-personalised advertisements recommended by AI based on personal preferences and behavioural data. If companies fail to respond to these structural changes, the revenue models of existing players that rely on mass advertising will rapidly lose their presence.

Direction for survival

In order for the advertising industry and media to survive, two approaches are required.

1. Experience design
   It will be important to create a system that naturally connects viewing and purchasing and is evaluated by algorithms.
   However, current social media and video streaming services are dominated by specific platform companies, raising the question of how much the existing advertising industry and media can penetrate this space. They will need to either move into existing social media networks and platforms and redesign their advertising and experiences, or build new platforms themselves and take the lead in experience design.
2. Developing customer IP
   Instead of the traditional model of simply buying ad space, preparing ads and selling them, companies will need to provide comprehensive support as partners to their customers, nurturing their brands, content and IP, and expanding them in the market.

In other words, the advertising industry will need to evolve from the traditional ‘advertising business’ to the ‘experience design business’ and ultimately to the ‘IP cultivation business’. The ability to use AI to design experiences and cultivate IP will be the key to survival.

Recommendation logic, purchase prediction, understanding brand momentum, price optimisation, promotion optimisation—what is needed to grasp the situation in real time and take action? Should these tasks be done manually, or should AI be used to support customers efficiently? In a future dominated by algorithms, it seems inevitable that those supporting distribution will also utilise AI. The players who survive will be those who control data and IP, can use AI and can redesign their relationships with users.

Conclusion

In this report, we touched on the evolution of AI, its benefits and its threats, and looked at how AI can potentially transform industry and society. We also discussed the future that AI may bring to the technology, media, and information and communications industries while highlighting recent specific trends.

As discussed in Chapter 2, the management agenda required in the age of AI will change drastically. The pace of change will continue to accelerate, especially in today’s rapidly changing social structure. To avoid being left behind, redefining your company’s value proposition will be an essential process. In other words, it will be important to identify the value that your company should create within the AI industry structure and then determine which areas to focus on to create a competitive advantage. We believe that the ‘industrial architecture’ framework discussed in this report, which organises the AI industry structure into three layers—social value, industrial value and functional value—will be helpful in making this determination.

¹ As examples, both the SDGs and the Paris Agreement were adopted in 2015 and came into force in 2016 (SDGs: United Nations Information Centre; Paris Agreement: Ministry of Foreign Affairs).

² Ericsson. Ericsson Mobility Report November 2025. Total mobile network traffic is expected to increase 2.4-fold, from 197 EB/month (2025) to 482 EB/month (2031). Fixed data traffic is expected to increase 1.9-fold, from 380 EB/month (2025) to 710 EB/month (2031).

³ Ministry of Economy, Trade and Industry, The Fifth Semiconductor and Digital Industry Strategy Review Conference.