#40 Society and Technology

The Australian government’s new regulation restricting social media access for individuals under 16 serves as a pilot initiative for governments worldwide engaged in similar debates. This law not only represents a measure aimed at protecting children’s safety but also signifies an attempt to redefine digital boundaries within national legal norms and test the state’s capacity to regulate technology platforms. Australia’s new legislation may mark a significant step in shaping the future of global technology policy.

In democratic systems, legislative processes concerning technology typically rely on lengthy technical consultations. However, in the case of Australia’s social media restriction, the process has progressed rapidly. The government led by Prime Minister Anthony Albanese can be said to have taken this step toward a social media ban through broad political consensus in response to growing public concerns and parental demands. This speed reflects the consequence of technology policy transitioning from a technical issue to a top-tier social agenda item.

The ban can be interpreted in several ways. While the emphasis on “under 16” may appear reasonable from a parental perspective, it can also be viewed as a serious anti-democratic restriction. According to The Wall Street Journal, platforms such as Reddit that have initiated legal proceedings against the Australian government argue that the ban seeks to exclude young people from the digital public sphere and restrict their political rights.

As you know, social media today is far more than a simple entertainment tool. It has become the primary medium through which young people acquire information on issues ranging from the climate crisis to global politics and develop political awareness. Critics argue that completely excluding this age group from the digital public sphere will deprive them of their right to democratic participation and freedom of access to information. Thus, this restriction risks creating a censorship mechanism designed to mold young people into “reasonable citizens” and control their political socialization processes.

Setting aside political debate, the practical requirements of implementation necessitate structural changes in the general functioning of the internet. The most reliable way to filter a specific age group online is to verify users’ ages without ambiguity. This implies the widespread adoption of “Age Assurance” systems. For social media platforms to comply with the law, they will need to further develop mechanisms to confirm users’ identities or ages. This process will disrupt the internet’s traditional anonymous structure and enhance its manageability and controllability. As a result, malicious and disinformation-containing content posted under anonymous identities will become subject to oversight. For states, this transformation will strengthen digital sovereignty while making it easier to hold malicious actors accountable. In addition, there are significant obstacles between the law’s intended goals on paper and the technical realities on the ground. Analysts highlight that such centralized restrictions may struggle against flexible digital tools and alternative workarounds.

Australia’s initiative will establish a global precedent regardless of how successfully the law is implemented in practice. Governments are increasingly signaling their willingness to assert sovereignty against international tech corporations. This regulation could serve as a reference for the United Kingdom and European Union countries that are considering similar legislation, as it expands the market for age verification technologies. If the model implemented in Australia gains public approval and overcomes technical barriers, we may witness the dawn of a new era in which internet access standards are fundamentally transformed.

Google’s latest research demonstrates that robots can now be trained without repeatedly testing in the real world. Simulations developed using videos generated by artificial intelligence can predict with high accuracy how well a robot will perform a given task. This finding signals a major paradigm shift in the field of robotics.

According to the research, a video-based simulator built on Google’s Veo technology can accurately predict robot performance in real-world tasks. Tests revealed a strong alignment between the simulator’s predicted success rates and the robots’ actual performance in the physical world. In practice, this means the simulator can reliably distinguish which configurations yield better performance.

Until now, robot training has been both expensive and slow. For a robot to reliably perform even simple tasks, it needed to repeat the same motion hundreds of times in the real environment. Each mistake meant loss of time, money, and hardware. Video-based simulations, however, shift most of this process into the digital realm. Companies can now test different scenarios virtually before producing physical robots and identify which settings will work effectively in advance.

It is no longer necessary to teach robots a task by having them perform it repeatedly in the real world. Instead, it is now possible to teach them “what happens if you do this” through videos generated by artificial intelligence. Moreover, this method can accurately predict how successfully the robot will perform the task in the real world.

The most critical outcome of this development is that companies can now conduct far more intensive and in-depth training in simulation environments. Confidence that the skills acquired in these environments can be successfully transferred to physical robots has also increased markedly.One of the fundamental challenges in the robotics sector has been access to the broad and diverse training data required for robots to reliably perform tasks.

While traditional approaches require robots to physically repeat tasks countless times, Google’s research shows that simulation- and data pipeline-based training achieves results far more quickly than methods relying solely on physical practice.

Silicon Valley has long discussed this approach, and Google’s robotics research has now provided it with strong scientific backing. This evidence is likely to accelerate global technology companies’ investments in robotics.

The emerging picture suggests that the key future need will not be producing more robots, but rather building powerful computing infrastructure to run these simulations. The primary driver of robot training is shifting from factory floors to data centers and computational capacity. In particular, GPUs are emerging as one of the most critical resources of this new era.

In conclusion, robot technology is undergoing a profound transformation in its impact on daily life. Robots are now learning by observing, not by experimenting. They are being trained first in digital environments, not in the physical world. This shift enables faster robot training but also raises broader questions about the future of production and labor.

We often overlook the environmental costs behind the technological services we use daily. The internet, cloud services, and digital applications create a sense of an “immaterial” world—as if they consume no energy, use no water, and generate no direct environmental impact. Yet nearly all sectors, from agriculture to mining and textiles to technology, rely heavily on water consumption. A t-shirt, a hamburger, or a smartphone requires thousands of liters of water across different stages of their production chains. Nevertheless, the environmental impacts of these sectors are generally normalized and even accepted as inevitable. Today, artificial intelligence stands as the most recent and controversial example of this invisible infrastructure. AI systems, distinct in their novelty and separation from other industries, clearly reveal the geographical and physical damage caused by their water usage.

Wired magazine’s article examining the debate surrounding AI and data center water use focuses on selective perception. The article shows that many sensational claims about AI’s water consumption are often based on incomplete or erroneous calculations, yet still generate a powerful social impact. The example of a data center in Chile, cited in journalist Karen Hao’s book and later found to contain a serious statistical error, illustrates how rapidly numbers can be transformed into symbolic narratives. As a result, the debate has moved beyond technical considerations to become a narrative positioning AI as an environmental threat.

Water usage by data centers is not uniform or equally impactful everywhere. Water consumption varies significantly depending on the geographic location, climate conditions, cooling technologies used, and the state of local water resources. In some regions, such facilities create serious pressure, while in areas with relatively abundant water supplies, their impact remains limited. Nevertheless, public debates often ignore these differences and rely on generalized and disproportionate rhetoric.

What is truly striking here is why AI has become such a central target. As the article notes, sectors like agriculture, textiles, and food consume far more water but do not face comparable public backlash. The reason lies in the fact that the outputs of these sectors—food, clothing, shelter—are socially perceived as more tangible and indispensable. Artificial intelligence, by contrast, remains a technological innovation whose benefits and necessity are still being debated relative to its societal contributions. Thus, the water debate has become an expression of a broader societal judgment about AI’s value. However, it is plausible to predict that as AI usage becomes more widespread and socially accepted, this issue of resource consumption will no longer dominate public discourse, much as it did not during the initial phase of the internet’s widespread adoption.

Ultimately, the issue of AI and water has moved beyond a mere environmental accounting problem. It is better understood as an expression of a broader value debate: which technologies we deem necessary versus which we question, and which uses of public resources we consider legitimate. After all, the central question regarding technology is rarely how much it consumes, but why and for whom it is consumed. In this context, the societal perception of something as a necessity becomes a crucial mechanism for tolerating large-scale resource use. Thus, as artificial intelligence establishes itself not merely as a functional tool but as a perceived societal need, its public perception may fundamentally change.