#41 Society and Technology

Merriam-Webster dictionary has named “Slop” its Word of the Year for 2025. But what exactly does this term mean, a concept we have increasingly encountered in recent times within the technology world and on social media? According to the dictionary’s definition, slop refers to “masses of content generated by artificial intelligence, typically low quality, careless, and filling our social media feeds”. Recently, we frequently encounter AI-generated images, short and long videos, and accounts built entirely around this type of content—both visual and auditory—which are now described using the term slop. With its entry into the dictionary, the word has given conceptual form to the confusion and unpleasantness we have long felt in our digital experiences but could not quite name.

The emergence of this term is a significant indicator of the structural transformation the internet has undergone. Once, the internet was seen as a new “public sphere”—an organic structure where people could freely debate under anonymous identities, create their own subcultures on forums, and center interaction around real human beings. But today, the monopolization created by social media companies and the growing volume of AI-generated content have drastically altered this picture.

It is precisely at this point that the “Dead Internet Theory,” long existing as an internet legend on the margins, has returned to public discourse with stronger arguments. The theory fundamentally claims that a large portion of internet traffic and content is no longer created by humans but by bots and algorithms. Where once there was communication between people, a vast sea of meaningless content—produced by algorithms solely to amplify interactions and generate profit—now dominates.

Slop is not merely a “spam” or advertising problem. It actually signals a fundamental shift within digital culture. The content presented to us no longer needs to have a purpose, a narrative, or an aesthetic concern. The goal is simply to fill the feed, hold attention for seconds at a time, and ultimately increase user dwell time on the platform. This transforms the internet from an active source of information shaped by its users into an endless pile of attention-grabbing nonsense.

Merriam-Webster’s choice may well be an indicator of growing awareness against this digital pollution. The widespread adoption of the term “slop” shows that users are no longer passively consuming whatever is placed before them and that the longing for meaningful, human-centered content is increasing. Therefore, in the coming period, the phrase “this content was not generated by artificial intelligence” may become a necessary condition for content to be regarded as credible by people.

As the role of AI chatbots in our lives expands, identifying their invisible risks becomes increasingly crucial. One of the most common commands we now encounter—“rewrite this as if a real human wrote it”—reveals the nature of our relationship with artificial intelligence. The intention behind this command is clear: to somehow delegate to the machine the output associated with human writing. But what does this tell us about language and thought?

A recent article published on Psychology Today, titled “When Sounding Good Replaces Thinking Well,” sheds light on how our expectations of AI have transformed into a relationship of appearance and self-perception. As is well known, the primary goal is to have AI-generated text rewritten again by AI to make it appear as if it were written by a human. Two questions arise here. The first is: what distinguishes this process from traditional editing? The second is: what would happen if we assumed the text were genuinely written by a human?

Regarding the first question, one could argue that for centuries our texts have already been edited and reviewed by others. So what distinguishes AI’s humanizing editing? Or does AI’s editorial intervention enhance the clarity and depth of the text?

It is undeniable that language forms the foundation of thought. Even without entering the debate over whether language is an absolute boundary for thought, a prerequisite for it, or its consequence, we can observe an unshakable bond between the two. Wherever thought exists, language is present; and wherever language exists, a “subject” is hidden. Even when reading a text whose author has long since passed away, we trace in the lines between the words their orientations, intentions, and the traces of thoughts that danced in their minds before being committed to paper.

The landscape we now face, however, undermines this ancient bond. AI-generated and humanized texts detach language from thought and from the author. These algorithms produce texts as if they were the product of a thought process that never existed. As the author rightly notes, even if all the lights are on, no one is home. In this sense, texts edited for the purpose of humanization differ fundamentally from traditional editorial revisions that enhance a text’s depth. In AI-generated texts, the author’s intention, depth, and polish vanish. As a result, our trust in the text—and by extension, in the author—is damaged, because the author has now disappeared.

Sounding human or producing rhetoric that mimics thought carries the danger of obstructing genuine thinking. Language, which is a representation of inner experience, is now being replaced by mechanical responses governed solely by social acceptance rules. In this process, the effort to be understood gives way to the desire to appear understood. The psychological consequence of this transformation is even more alarming: the issue is no longer about feeling good, but about appearing good.

At this point, we are forced to choose between appearing good and thinking deeply. If aesthetic and intellectual refinement are no longer the product of genuine effort, our capacity to discern will inevitably atrophy. The trust implicitly granted to texts we assume were written by humans is now being shaken by machines that merely simulate humanity.

Thus, although commands are designed to humanize texts, the process itself is increasingly detached from humanity. Rather than rapidly humanizing the artificial, we must bend toward the human and avoid losing it.

Numerous projections about the future of artificial intelligence in the workplace have long been debated. Some extend this transformation over decades, while others argue a profound rupture will occur within a much shorter timeframe. This second view has become more visible recently due to statements by Charles Lamanna, a senior executive at Microsoft responsible for Copilot and corporate AI agent strategy. One such projection suggests that artificial intelligence will transition from its “assistant” role to become an active agent that directly performs actions and end-to-end processes within the next six months to six years. This is not merely a technical advancement in how work is done—it implies a fundamental restructuring of organizational structures and even job definitions.

Until now, corporate AI solutions have largely occupied a supportive role, accelerating human tasks such as summarizing emails, preparing reports, and classifying customer inquiries. These are systems known as “agent-based AI.” Unlike tools that simply respond to commands, these AI agents can independently plan and execute tasks by accessing tools, corporate data, and software, guided by predefined goals. They can loop back to humans when necessary and sustain processes autonomously. In the new era, the emerging “agent-based” AI approach is expected to rapidly surpass these boundaries. AI agents now operate autonomously within critical business processes by connecting to tools, corporate data, and organizational knowledge. Their ability to remain continuously active, escalate issues to humans when needed, and produce results at scalable speed transforms them into digital “employees.”

The effects of this transformation are already observable across operational domains, from finance to customer service. In processes such as invoice processing, inventory management, and end-to-end resolution of customer requests, AI agents are increasing accuracy while significantly reducing manual workload. IDC data shows that companies adopting these systems early are gaining not only cost advantages but also faster innovation and revenue growth. This trend demonstrates that AI is no longer merely a tool for efficiency but has become a strategic layer directly shaping how companies compete. Organizations integrating agent-based AI into their workflows can make faster decisions with lower operational costs, while those left behind must rely on greater human resources and budgets to perform the same tasks. For example, a company using AI agents to manage customer requests end-to-end can offer 24/7 service while reducing costs, whereas competitors lacking this infrastructure face higher operational expenses and fall behind in customer experience.

This organizational transformation is also restructuring information and office work processes. Routine and repetitive tasks are increasingly delegated to AI, while human employees shift toward roles in oversight and strategic decision-making. This process also brings new professions into existence, such as agent developers, AI strategists, or digital worker managers. However, this transformation is neither unlimited nor uncontrolled. For the transition to be sustainable, it is critical that AI agents have only the necessary access rights, and that these permissions are continually redefined as their responsibilities grow.

In conclusion, the future of artificial intelligence in the workplace appears to be moving beyond its assistant role. The key question is how quickly and how consciously organizations will respond to the agent-based AI era expected to emerge within the next six months to six years. This transformation is poised to dominate the agendas of both employers and workers.