AI and Anthropocentrism / Transition to Industry 5.0 and Reconstruction of the Legal System: Random Thoughts on Moving from Efficiency to Resilience and Anthropocentrism

0 Introduction

Over the past decade or so, the manufacturing industry has been pushing forward with digitalization and automation under the banner of Industry 4.0. Connectivity via IoT, big data analysis, and autonomous production control via AI. The unwritten rule behind this was to eliminate the uncertain element of humans from the production process as much as possible and to pursue thorough efficiency and cost reduction.

However, market signals have changed dramatically. According to the Industry 5.0 Trend Report published by StartUs Insights and the analysis by Andreas Neidhart, the tide of the industry is shifting significantly toward Industry 5.0. The focus is shifting from machines replacing humans to machines cooperating with humans.

Industry 5.0 does not deny the pursuit of efficiency, but it shifts its center of gravity toward human-centricity, resilience, and sustainability. In this article, I will attempt to organize my random thoughts on how this paradigm shift forces changes in conventional legal systems and legal thinking, using the startup trends and key trends of 2,778 companies analyzed by StartUs Insights as material.

1 The Location of the Problem: The Shift in Values and the Rigidity of Legal Frameworks

In discussions about law and technology, the main point of contention has been how full automation conflicts with legal systems. The trolley problem in autonomous driving, the ethics of autonomous weapons, and copyright by AI. There, the void of rules in areas where humans are absent has been questioned.

However, the vision presented by Industry 5.0 is different. The human-centric approach shown in the StartUs Insights report assumes that AI and robots are positioned as human partners and are closely involved both physically and cognitively. As Neidhart points out, the source of value is shifting from cost reduction through automation to value creation through resilience, sustainability, and hyper-personalization.

From a legal perspective, this shift toward collaboration and resilience contains more troublesome problems than full automation. Law traditionally emphasizes predictability and legal stability, and tends to fix existing responsibility allocations. How can current legal frameworks optimized on the premise of efficiency—for example, contract practices with strict performance deadlines, routine occupational health and safety legislation, and product liability laws premised on standardized products—be consistent with a flexible, human-centric new industrial structure? This is the central question I will examine in this paper.

2 The Location of Responsibility in Human-Centric AI and Collaborative Robots

At the core of Industry 5.0 are cobots that work in the same space as humans and human-centric AI that complements human decision-making. Among the impacts of the trends analyzed by StartUs Insights, Human-Centric AI accounts for 27% and Cobots for 13%, with the two together making up 40% of the total.

The general principle for conventional industrial robots was to be physically isolated from humans by safety fences, based on laws and regulations such as occupational health and safety rules. The law ensured safety through isolation. However, cobots in Industry 5.0 remove this physical boundary. Cobots like those from RBW in South Korea and Genitor in the US, introduced in the report, use proximity sensors, LiDAR, and AI vision to recognize human gestures and adjust their movements.

What emerges here is the complexity of legal liability when an accident occurs. If damage occurs while machines and humans are working in close cooperation and interaction, it becomes extremely difficult to distinguish whether it is a machine defect or a human operational error.

For example, if an accident occurs after AI proposes an efficient action and a human approves and executes it. Is the AI's proposal itself a defect, or is the judgment of the human who approved it negligent? Or, if the AI predicted danger and issued a warning, but the human could not correctly interpret the meaning of that warning. Can that be called a flaw in the design of the human interface?

Human-centric means reserving final decision-making and responsibility to humans. However, when systems become more sophisticated and reach a level where AI estimates human intent and acts, the question of whether humans can maintain substantial control—that is, the effectiveness of Meaningful Human Control—will be sharply questioned in the areas of labor law and tort law.

The report introduces cases such as the UK's SiteAssist, an AI platform that supports compliance and safety checks in high-risk environments such as construction sites. While it supports the fulfillment of the duty of care for safety, it also means that workers' biological data and work efficiency are monitored in real-time through wearable devices and edge analytics. While useful from the perspective of ensuring worker safety, excessive monitoring can lead to violations of privacy rights and psychological stress from being constantly evaluated by AI. The law needs to act as a strict monitor regarding whether human-centric in Industry 5.0 truly refers to human well-being or to the optimization of humans as a resource.

(Reference)

3 Resilience and Agile Supply Chains: Contract Practice and Economic Security

Another pillar of Industry 5.0 is resilience. With the manifestation of pandemics and geopolitical risks, just-in-time supply chains that pursued efficiency have exposed their vulnerabilities. In the report as well, digital sovereignty and cybersecurity are cited as key trends along with resilient and agile supply chains.

From a legal perspective, this shift in values affects corporate governance, especially the interpretation of directors' duty of care. Conventionally, reducing inventory to the limit and selecting the lowest-cost supplier was considered a rational management decision. However, now that resilience is emphasized, even if costs increase, investing in the diversification of suppliers and the visualization of supply chains will be required as the fulfillment of the duty of care.

Startups like Norway's Smartcube Solutions and the US's CommonShare are realizing data sharing and traceability across the entire supply chain. The use of such platforms creates a new issue of how to balance contractual confidentiality obligations with ensuring transparency. Schemes that share inventory information and alternative production capacity with competitors in an emergency create a tense relationship with antitrust cartel regulations, so careful legal design is required.

Furthermore, the digitalization and decentralization of supply chains lead to an increase in cyber risks. The report mentions that industrial ransomware attacks have increased by 87% year-on-year. From a legal perspective, this goes beyond mere security issues and is deeply linked to the context of national security, such as data localization regulations and the Economic Security Promotion Act.

Data sovereignty infrastructure provided by startups like the UK's Valarian helps companies legally control where their data is stored and who can access it. This resonates with discussions surrounding the balance between exclusive rights to data and access rights, as seen in Europe's GDPR and Data Act. Corporate legal departments are being forced to go beyond mere compliance to strategically design data location and management authority, thereby protecting digital sovereignty from geopolitical risks.

4 Hyper-personalization and the Loss of Product Liability Law Standards

Industry 5.0 accelerates the transition from mass production to mass customization, and even to hyper-personalization. According to the report, one in five consumers is willing to pay a 20% premium for personalized products. This is made possible by 3D printing and digital twin technologies.

Examples such as the German company goatAM, which uses 3D printers to manufacture customized drugs tailored to each patient's condition, and CTO models like Mexico's Prima, fundamentally shake the premises of product liability laws and product safety regulations.

Conventional legal systems have operated on the premise that standardized products are manufactured and distributed in large quantities, treating deviations from those standards as defects. However, for unique, one-off products, the standards to be referenced are generated each time.

When a drug with different ingredients or shapes for each patient is output by a 3D printer, who is responsible if health damage occurs? Is it the printer manufacturer, the algorithm designer of the control AI, the doctor or pharmacist who entered the parameters, or the patient themselves who provided the data? In this domain, where product defects and service errors are inextricably linked, it becomes difficult to maintain a balance between victim relief and innovation using only conventional product liability jurisprudence.

To address this, in addition to inspecting the final product, process certification that guarantees the reliability of design algorithms and manufacturing processes, and ensuring complete traceability of manufacturing history using blockchain or similar technologies, may come to function as the basis for legal exemption requirements or liability claims.

5 Normalization of Sustainability and the Circular Economy

In Industry 5.0, sustainability is no longer a 'nice-to-have' element but is placed at the center of operations. Green technology and circular economy principles are becoming essential not only for cost reduction but also for fulfilling ESG obligations.

Solutions such as the industrial waste upcycling handled by the UK's Cocoon and the carbon capture technology developed by the US's Mitico are in line with the tightening of environmental laws and regulations. In a legal context, this is a flow from soft law to hard law. Environmental considerations, which were previously part of a company's voluntary CSR activities, are being elevated into clear legal obligations through mechanisms such as the European Corporate Sustainability Due Diligence Directive and the Carbon Border Adjustment Mechanism.

This is where AI and law intersect in the verification of environmental value. Emission monitoring and energy management systems using edge analytics, as mentioned in the report, are taking on the character of evidence-generating devices for regulatory compliance, rather than just energy-saving tools.

Conversely, environmental claims without such technical backing will increasingly invite legal risk in the future. This includes violations of the Unfair Competition Prevention Act and consumer laws due to greenwashing. The visualization of environmental impact through technology narrows the room for companies to argue that they 'did not know.' Legal departments must verify the authenticity and evidentiary value of AI-generated environmental data and fulfill their accountability to regulators and stakeholders.

6 Reflections

The worldview of Industry 5.0 depicted in the StartUs Insights report is not a dystopia where technology leaves humans behind, but a future where technology enhances human resilience and harmonizes with the environment. However, for this to be realized, the legal system must also evolve from the efficiency-oriented and rigid system of the Industry 4.0 era to a human-centric and flexible one for the 5.0 era.

Responsibility sharing in human-AI collaboration, sovereignty disputes over supply chain data, safety standards for personalized products, and legal commitments to sustainability—these are not merely technical challenges, but also legal-philosophical questions about what kind of social order we desire. The discussion surrounding AI and law should also move from the phase of how to regulate AI to the phase of how to collaborate with AI and how to design laws to protect the values of human society.

(Magazine) "Reflections on AI and Law"

*See below for the table of contents

note General Terms of Service Article 3, Paragraph 2, First Sentence
3.2 The copyright of digital content created by the creator belongs to the creator.