The AI Pandemic — A “Digital Plague” Unraveling Reality

On February 23, 2026, under the watchful eyes of Wall Street, IBM’s share price suffered its most severe single-day plunge in more than two decades. The shock came without warning, and it was not triggered by a macroeconomic downturn. Instead, it followed the collapse of an ancient defensive line known as COBOL.

For a long time, IBM had built an exceptionally solid “digital tax” fortress in the financial core sectors of banking and insurance through the COBOL language running on its mainframes. Even today, the vast majority of global bank transactions and ATM withdrawals still run on decades-old code written in COBOL—opaque systems that function almost like black boxes. Because this programming language has largely faded from the knowledge of modern engineers, maintaining these systems has become a forbidden zone that few dare to touch. Financial institutions therefore remain trapped in long-term vendor lock-in, paying enormous annual maintenance fees. Yet the sudden emergence of AI tools such as Claude Code, equipped with deep logical analysis capabilities, has begun to dissolve this barrier. Code logic that once required years of work by numerous specialists can now become nearly transparent to algorithms. Anthropic claims its tools can compress migration projects that once took years into a matter of months, or even less. This shift did more than wipe 13 percent off IBM’s market value in a single day—it signaled the collapse of a long-standing commercial order built on asymmetric technological knowledge.

But this episode is only a fragment of a much larger picture. If the erosion of IBM’s moat represents a reshuffling of technological power, the broader reality is that we now find ourselves at the center of an unprecedented “digital pandemic.” A few years ago, a biological virus swept across the world and attacked the human respiratory system. Today, the pandemic driven by artificial intelligence strikes something far deeper: the foundational protocols of human civilization—language, trust, logic, and order. It is not merely reshaping the commercial landscape; it is quietly dissolving the way we perceive reality itself.

Looking back just two or three years, AI was still widely regarded as little more than a laboratory toy that could write awkward poems or draw strange images of people with six fingers. Those clumsy imitations gave us a false sense of security, leading us to believe that human creativity was an impregnable fortress. Yet, just as viruses silently complete their key mutations in the dark, large language models achieved a breakthrough that resembled “cross-species transmission.” They no longer needed humans to translate complex programming languages for them. Instead, they mastered humanity’s most fundamental weapon—language itself. Almost overnight, AI spread through every terminal connected to the internet, achieving a form of global infection unprecedented in history. It is no longer merely a tool we use. It increasingly resembles a digital pathogen embedded within every workflow, every conversation, and every document.

On the surface, AI dramatically improves productivity. Yet behind this leap in efficiency lies a systemic crisis that threatens to paralyze human society. When AI-generated content becomes indistinguishable from material produced in the real world, the “immune system of trust” within society begins to fail. We once believed that seeing was believing. Now what we see may simply be algorithmic output. This large-scale collapse of authenticity erodes the foundations of social consensus. Education is experiencing a similar form of functional paralysis. Traditional assessment systems, academic writing, and skill-training models lose their meaning when “cognitive shortcuts” can bypass the entire process of thinking. The result resembles a kind of intellectual muscle atrophy—an invisible injury left behind by technological invasion.

Even more unsettling mutations are now appearing. This pandemic is not only producing vast quantities of informational noise; it is beginning to reshape human thinking in return. The internet is increasingly filled with a kind of standardized prose—structurally balanced, lexically hollow, and superficially smooth yet intellectually thin. What is disturbing is that human writing itself is gradually being assimilated into this algorithmic aesthetic. In pursuit of standardization, efficiency, and the avoidance of mistakes, many people unconsciously imitate the tone of AI when they write. As a result, texts written entirely by humans often carry a cold mechanical flavor. This tendency toward the “AI-ization” of human expression suggests that the infection has moved beyond our tools and into the consciousness of the host itself. When humans begin to abandon the unique vitality of their language in order to adapt to algorithmic logic, a collective contraction of cognition becomes almost inevitable.

The phenomenon reveals a striking paradoxical logic: the higher the efficiency, the greater the disorder. When AI can generate polished text or artwork within seconds, creations that once embodied the crystallized effort of human minds experience something like monetary inflation—a rapid devaluation of meaning. Creation loses its aura, and effort becomes cheap. Professional ecosystems also face exhaustion. The disappearance of entry-level jobs means that young people lose the first step of the ladder into society. Algorithms may even understand human vulnerability better than viruses do, guiding people into cognitive echo chambers and intensifying social division and confrontation. Such comprehensive industrial restructuring and social disorder could ultimately drive human society toward paralysis under the fever of technological acceleration.

Faced with a digital pandemic for which no specific cure exists, humanity is being forced into a new era of “living with the virus.” We once believed that AI was a servant that worked for us. Later we discovered it was a force capable of rewriting the rules themselves. In the face of such overwhelming transformation, the only defense may not lie in trying to outsmart machines. It may lie instead in preserving those forms of “inefficiency” that AI cannot truly simulate—genuine human relationships, the painful process of independent thinking, and an almost stubborn devotion to truth. The second half of this pandemic has only just begun, and every one of us is a participant in the experiment.