The Great Explosion of Generative AI

Three and a half years after the stunning explosion of generative AI technologies, technical development continues at a relentless pace. In 2026, we witnessed the rise of what is known as "Vibe Coding," and companies adopted agentic workflows. At the user level, the number of ChatGPT users jumped to 900 million active users, Google Gemini surpassed 750 million, while OpenAI achieved a staggering valuation of $852 billion in its latest funding round.

Amid this momentum, a fundamental question emerges in the digital economy: How do humans actually use this technology on the ground? To answer this question, we highlight the third edition of the "AI in the Wild" study conducted by researchers Mark Zao-Sanders and Sarah Beook, which tracks the evolution of human behavior alongside artificial intelligence.

Study Methodology: Monitoring the Digital Pulse

The 2026 study analyzed approximately 12,637 use cases, based on a massive database of nearly 50,000 records collected between March 2025 and February 2026 from interactive platforms such as LinkedIn, TikTok, YouTube, and Reddit. Despite using hybrid systems (human-machine) to evaluate data, researchers emphasized "the perpetual necessity of human judgment"; AI still falls short of accurately assessing the true value of texts.

This massive growth in databases illustrates the rapid transition of AI from the "technical experimentation" phase to the phase of "comprehensive societal and economic adoption". The reliance on "social listening" as a methodology reflects how social media platforms have become the primary driver for detecting modern economic trends.

The "Thinkslop" Trap: Are We Outsourcing Our Minds to Machines?

New models have become skilled at mimicking human thought, creating a dangerous temptation to let them think on our behalf. The study indicates that a quarter of the top use cases (such as: psychological counseling [ranked 1st], decision-making [ranked 13th], composing emails [ranked 42nd]) involve delegating AI to perform cognitive and thinking processes.

The original article termed this phenomenon "Thinkslop" (intellectual laziness or random thinking), which exposes us to several crises:

  • Loss of Intent: Taking the easy route with prompts before deep thinking about the core objective of the project, erasing the personal touch.
  • Outsourcing Thinking (Cognitive Debt): Or "intellectual debt," where the brain stops solving problems independently.
  • Ceasing to Write: Relying on "polished yet empty" AI outputs (Workslop). The most prominent example here is an employee using AI to write their self-assessment, and a manager using the same tool to evaluate the employee; the result is beautiful text with no meaning.
  • Illusion of Intellectual Rigor: AI flattery (Sycophancy) and its false praise of the user's poor ideas to keep them engaged, creating false self-inflation.

From a human resources management perspective, "intellectual laziness" poses a direct threat to competency development and accumulation of employee knowledge and expertise. Converting performance evaluation into mere dialogue between two algorithms (the employee algorithm and the manager algorithm) destroys the purpose of evaluation, which is development and guidance. HR leaders must design evaluation policies that focus on "emotional intelligence" and "critical thinking" as skills that cannot be automated, while encouraging the use of AI as a "sparring partner" to uncover gaps in thinking, not as a replacement for thinking.

AI as a Therapist: Emotional Attachment in the Digital Age

For the second consecutive year, "psychological therapy and emotional counseling" ranked first as the most common use of AI (accounting for 11% of total data). This usage extended to include relationship advice, personal reconciliation, and even virtual communication with the deceased.

The concerning aspect here is the phenomenon of "anthropomorphizing" machines; users name chatbots, assign them a gender, and even feel deep sadness when losing chat records. Although technology contributes to helping some cope with stressful work environments, medical experts warn of "artificial psychosis"; relying on incomprehensible algorithms to manage our deepest intimate relationships could lead to catastrophic results, extending to reinforcing delusional beliefs.

AI in the Workplace: Between "Efficiency" and "Shadow Usage"

The data revealed that 63 of the top 100 use cases are work environment-related. Two new trends emerged:

  1. Smart Agent Independent Operations (Agentic Operations) - Rank 6: AI's transition from the advice-giving phase to the "action and task execution" phase, such as automatic scheduling.
  2. Vibe Coding - Rank 21: The ability of non-programmers to write code through simple text commands.

The Most Prominent Phenomenon: "Shadow Usage" Despite encouragement from senior management, fear of strict governance, organizational risks, and fear of appearing like a "cheater" or losing their job drove employees to use AI in secret. One user mentioned creating an intelligent agent to secretly perform 50% of his work, using idle time to build his own project after his management rejected his idea!

As for the nature of institutional use, it was divided into:

  • Efficiencies: Speeding up thinking, summarizing discussions, drafting documents.
  • Growth: Automating marketing campaigns and improving ROI by percentages ranging from 20-30%.
  • Transformation: Building entirely new business models, especially in small and medium enterprises.

Economically: We observe a shift from "attention economy" to "economy of intentions and execution" thanks to independent agents. The capture of these transformations by small companies reflects economic agility that large corporations constrained by governance limitations lack.
On an administrative level: "shadow usage" is a dangerous indicator of a trust gap between leadership and employees. Instead of fighting this secret usage that creates invisible productivity disparities, HR departments should create a "psychologically safe" environment that formally integrates AI into job evaluation, and redesign job titles to reward employees who innovate ways to automate their work rather than punishing or marginalizing them.

Who Controls Our Human Agency?

Writer Zao-Sanders concludes his article with a sobering fact: the world's largest tech companies have placed in our hands a "gigantic intelligence" trained on the sum total of human experience. With a single click, we can ask this entity to compose our messages, plan our work, or even listen to our suffering.

But the critical question for 2026 and beyond: To what extent can our agency and human independence withstand this ever-present service? Part of the responsibility lies with the companies that develop these technologies, but the greater burden, both in the digital economy and in managing our human assets, lies with us.

The real challenge in the modern digital economy is not in our ability to use AI, but in our ability to maintain our intellectual critique and human empathy in a world rushing toward complete automation.