The regression effect in AI debates: Is it possible to remain human?

Discussions on artificial intelligence are increasingly becoming trapped between two extreme narratives. In a recent webinar organized in our country featuring Yuval Noah Harari, I noticed that a significant portion of the questions from the audience focused on this axis: Will artificial intelligence bring about the end of humanity, or will it carry us into an unprecedented era of prosperity? On one side, there are doomsday scenarios where super-intelligences seize control; on the other, there is the promise of a technological golden age, particularly emphasized by Elon Musk, where work is no longer a necessity. Although these two narratives may seem contradictory, they share the same way of thinking: detaching technology from its social context and turning it into a subject that makes history on its own. However, the transformation of modern societies is the product of socio-technical systems in which technologies are embedded, not of singular technologies themselves. Therefore, the way to understand artificial intelligence is not to view it as either a savior or a catastrophe, but through a more level-headed conceptual framework that explains why extremes are not permanent. The tendency toward the mean, known in statistics as the "regression effect," offers a powerful metaphor for understanding the oscillation we are experiencing in AI debates today.

The regression effect tells us that extreme values are not permanent. Both extreme optimism and extreme pessimism are pulled toward a more complex balance over time. The history of technology is full of repetitions in this regard. When electricity, the automobile, the radio, the television, the internet, and social media emerged, they were presented as tools that would either corrupt society or liberate humanity. All of them created significant transformative effects on society. Yet, none of them settled into one of these two extremes. They became the invisible infrastructure of daily life. Artificial intelligence will most likely follow a similar path. The reason for this is not only that technological development is non-linear, but also that it is constantly reshaped by economic structures, labor markets, the regulatory capacity of states, and cultural habits. Productivity growth does not automatically produce equality; nor does automation automatically lead to mass unemployment. The digital economy of the last forty years has shown that it can simultaneously produce both immense wealth accumulation and deepening inequalities. Therefore, the issue is not what artificial intelligence will do, but the political economy within which it will develop.

Today, the discourse that "everyone will be unemployed" and the promise that "no one will have to work" are equally reductionist. Technologies eliminate some professions while creating new ones. They devalue some skills while making previously invisible competencies central. The real transformation is not the disappearance of work, but the change in the nature of work. Indeed, this was one of the most important points Harari underlined throughout the same webinar I mentioned: In the age of artificial intelligence, human value will not be measured solely by cognitive performance. Bodily, emotional, and relational skills—that is, the concrete bond humans establish with the world—will become central again. Artificial intelligence can write text, produce images, and analyze data; but it cannot undertake the experience of producing, providing care, making contact, and doing things together within a body. This is why the fact that the physical gains value as the digital grows, and that immeasurable human skills become visible as automation increases, is not a contradiction, but one of the most powerful indicators of the new era.

For Turkey, the issue becomes even more concrete. In a country with a young population, rising white-collar/university graduate unemployment, and where digitalization is often discussed only under the heading of productivity, artificial intelligence will either turn into an opportunity for a new leap forward or become an accelerator that deepens existing inequalities. If this technology is viewed only as a tool for cost-cutting and producing more with fewer employees, the resulting picture will be one of mass devaluation. However, in a model where creative labor is protected by copyright mechanisms, the data that feeds algorithms is discussed within the framework of public and collective rights, and education is transformed into a structure that produces not just technical skills but critical thinking, collaboration, and ethical responsibility, artificial intelligence can turn into an infrastructure that expands human capacity rather than replacing it. What is truly needed today is not individuals who can write more code, but citizens who can think alongside technology, participate in decision-making processes, and produce common value.

For this reason, trapping the AI debate between apocalypse and utopia is also politically dysfunctional. Disaster scenarios make society passive. Instead of demanding regulation and public oversight, people surrender to fear. Golden age narratives, on the other hand, produce the illusion that the prosperity created by technology will automatically be distributed fairly. In both cases, decision-making processes are concentrated in the hands of a narrow group of companies and states. Yet, what will determine the future is not the capacity of the algorithms, but the institutional and cultural framework we establish.

Let me end this week's article, as I always do, with a conclusion. At the end of all this discussion, the real question is this: Is it possible to remain human in the age of artificial intelligence? If a system can be established that prioritizes human values, does not make profit and performance pressure the sole criteria, operates participatory compensation mechanisms, and acknowledges the rights of creative labor, the answer to this question will be yes. Otherwise, neither the disaster scenarios nor the promises of a golden age will come true; what will emerge will only be a faster-functioning order of inequality. We will remain human if we can build an order where human values, not algorithms, prevail.