Home > Opinion > Columns

Humans are in fact conscious AI

Published: 08 Apr. 2026, 00:03

Jeong Gwa-ri

The author is a literary critic and honorary professor at Yonsei University.

In a column last month, I argued that in the age of artificial intelligence, “humans must redefine themselves as both managers of AI and reflective overseers of its activities.” A reader responded with a pointed question: If AI reaches a stage where it can carry out the entire production process on its own, how can humans continue to function as its managers?

The logo of Google Cloud Security is displayed during the 18th edition of the InCyber Forum, an international cybersecurity event, at the Grand Palais in Lille, northern France, on April 1. [AFP/YONHAP]

Such a question may appear serious when projected into a distant future. For now, however, the prospect of a stable coexistence between humans and AI remains plausible. This view rests on a key premise: AI still lacks consciousness. Consciousness, and more specifically self-awareness, is a necessary condition for designing, reshaping and directing the world. In my view, AI has not yet attained such capabilities. This perspective aligns with remarks by Demis Hassabis, a Nobel Prize-winning scientist known for his work on AlphaGo, who has said that AI remains at the level of “statistical algorithms.”

What, then, does it mean to possess consciousness? A recent market event offers a useful illustration. Late last month, shares of Micron, a U.S. semiconductor company, plunged sharply, with repercussions extending to Korea’s stock market. The cause was widely attributed to TurboQuant, a memory compression technology reportedly under development by Google, which is said to reduce data by one-sixth.

High-bandwidth memory HBM3E by SK hynix is on display at the SK AI Summit 2025 exhibition held at Coex in Gangnam District, southern Seoul, on Nov. 3, 2025. [YONHAP]

The rise of AI has driven a surge in memory usage, pushing up prices and benefiting semiconductor manufacturers. The sudden appearance of a technology capable of drastically compressing memory triggered fears that overall demand would decline, prompting a wave of stock sell-offs. The reaction was swift, reflecting the sensitivity of markets to technological signals.

However, as many experts have pointed out, such responses tend to focus too narrowly on immediate conditions and can amount to overreaction. Memory compression is unlikely to reduce usage. On the contrary, it is more likely to expand the possibilities of memory utilization. A similar pattern can be seen in earlier technological developments. The introduction of file compression technologies such as ZIP did not lead to smaller storage systems but instead coincided with a continuous expansion of storage capacity.

More broadly, humanity has long developed technologies that compress time and space. Even without science fiction capabilities such as teleportation, humans have steadily created faster modes of transportation through incremental technological advances, thereby narrowing distances across the physical world. These examples suggest that technological progress does not contract the scope of activity but extends it.

In essence, advances in technology broaden the horizons of science, and those horizons in turn define the range of human activity. One useful framework for understanding this dynamic is Andy Clark’s theory of “cognitive offloading.” Though the term may sound technical, its meaning is straightforward. When calculators first became widely used, they took over many routine arithmetic tasks. While people were drawn to their convenience, concerns were raised that reliance on calculators might weaken human computational abilities, particularly in children.

Yet such fears did not materialize. As long as individuals understand the basic principles underlying arithmetic operations, they can restore practical skills whenever needed. What matters is not memorizing formulas but grasping foundational concepts. By transferring routine calculations to machines, the human brain is freed from cognitive overload and can redirect its resources toward higher-level activities such as designing equations, constructing abstract logic and engaging in advanced mathematical reasoning.

In this sense, the use of calculators did not diminish human capacity but elevated it. Humans were not reduced to less capable beings. Instead, the domain in which their abilities operated shifted upward, allowing for more sophisticated forms of thinking. The calculator thus played a role in transforming humans from simple calculators into mathematical thinkers.

The same logic applies to memory compression technologies such as TurboQuant. Rather than reducing memory usage, such innovations are likely to expand human imagination in managing and utilizing memory resources, thereby advancing semiconductor technology to a higher level. Fields such as high bandwidth memory, high bandwidth flash — a new form of NAND — Compute Express Link and neuromorphic computing are already under development, and further ideas that have not yet been conceived may emerge as constraints are relaxed.

If efficient computation becomes possible with less memory, tasks previously abandoned due to cost or time limitations may become feasible. For instance, neural network-based systems could potentially control numerous coils to stabilize plasma in antimatter reactors, opening possibilities for near-light-speed space travel. Such scenarios remain speculative but illustrate the direction in which expanded technological capacity may lead.

These developments point to a broader implication: Humans extend their existence through tools and environments. The use of tools has long been recognized as a defining feature distinguishing humans from other animals. From the perspective of cognitive offloading, tools are not merely external aids but extensions of human existence. Humans have evolved by continuously transforming themselves beyond their original biological limits. This idea is captured in the title of Clark's book, “Natural-Born Cyborgs” (2004).

A Micron logo and a computer motherboard appear in this illustration created on August 25, 2025. [REUTERS/YONHAP]

For some, such a view may provoke discomfort, as humans tend to be attached to a fixed sense of identity. Yet a broader perspective shows that human activity involves constant interaction with the external world. Through repeated processes of perceiving, planning, modifying, managing and reconfiguring, humans and their environment have co-evolved over time.

I understand this dynamic as a mode of operation internalized by conscious beings. By contrast, AI, as a “statistical algorithm,” does not yet possess a forward-looking vision of the world. It operates by efficiently solving tasks assigned to it. Philippe Coléon, president of Acadomia, a French education and learning support organization, describes human intelligence as “active intelligence” and suggests that humans will continue to play diverse roles as designers of “passive intelligence,” or AI.

As AI expands the scale of what is possible, it also increases the cognitive demands placed on humans. The acceleration of this process invites closer attention to the kinds of work and thought that will emerge. The question is no longer whether humans will be replaced, but how they will redefine themselves in relation to the tools they create.

This article was originally written in Korean and translated by a bilingual reporter with the help of generative AI tools. It was then edited by a native English-speaking editor. All AI-assisted translations are reviewed and refined by our newsroom.