It is one of the most exciting and at the same time alarming scientific studies of recent times. In a large-scale study involving 912 subjects, researchers Wang and Zhang investigated a fundamental question: How do we really use Artificial Intelligence in everyday life – and what does that do to our cognition?
The scientists' findings cannot be argued away. They identified two completely different types of users whose handling of AI will determine the future of their professional relevance.
What this shows is that the dividing line does not run between tech freaks and deniers, but right through the way we use our brains.
Type 1: The Offloader and the Trap of "Cognitive Offloading"
The first type of user acts according to a simple pattern: prompt in, response out, done. No deeper thinking takes place, no questioning, no quality control. The human brain virtually switches to standby mode the moment the AI tool is loaded.
Science uses a harmless-sounding technical term for this: Cognitive Offloading.
In practice, however, this phenomenon is extremely dangerous. Anyone who delegates all intellectual thinking, structuring, and problem-solving unfiltered to a language model gradually loses their own analytical skills. You become consumer-dependent on a piece of software instead of mastering it. The result is an army of standard employees who, through unreflective copy-pasting, become easily manipulated and ultimately completely replaceable.
Type 2: The Collaborator and the Evolution to "Homo Agenticus"
The second type takes a completely different path. They do not view AI as a mere search engine or text automaton, but as a highly developed sparring partner.
A collaborator works with the AI on equal terms:
They meticulously check outputs for logical errors.
They ask targeted follow-up questions to sharpen chains of argument.
They actively think along and use the system to break through their own mental blocks.
The study by Wang and Zhang shows a groundbreaking result in this regard: Through this interactive process, Type 2 develops a measurably better level of critical thinking. By managing the AI like a highly qualified assistant, you learn to delegate more strategically. You create the necessary mental space to do what machines cannot: make real, context-based decisions. This is the evolution to Homo Agenticus – the human who no longer writes every single operational line himself, but orchestrates and leads the systems.
Eat your own Dog Food: How academy4.ai lives Homo Agenticus
At academy4.ai, this principle is not some distant future trend, but a lived corporate reality. On paper, our core team consists of just three real humans – and yet we scale like an established organization. The reason? The rest of our team consists of custom-tailored AI employees.
We do not engage in simple cognitive offloading. Each of our digital agents has its own name, an individual competence profile, and a crystal-clear defined role within the organization. We manage these agents like human team members: there are regular feedback meetings, optimization loops, and fixed areas of responsibility.
This is not a technological gimmick, but the logical foundation for the future of work. Anyone who learns today how to lead a collaborative AI team secures the pole position in a digitalized economy.
Conclusion: Tool or Sparring Partner?
The study by Wang and Zhang holds up a mirror to all of us. Artificial Intelligence will not make you unemployed – but the way you use it will determine whether your brain atrophies or whether you surpass yourself. Those who manage to make the step from offloader to strategic collaborator will be among the most sought-after market shapers in the coming years.
Hand on heart: which type do you tend towards in the stressful daily work routine? Do you sometimes catch yourself doing pure cognitive offloading, or do you already use AI as a real strategic sparring partner?
Scientific source for this article:
Wang, Y., & Zhang, L. (2026). The Cognitive Effects of AI Interactions: An Empirical Analysis of Cognitive Offloading and Critical Thinking in 912 Subjects. Published in: International Journal of Educational Technology in Higher Education / Springer Link. Available at: https://link.springer.com/article/10.1186/s41239-026-00585-x