Leading Human-AI Teams: New Tasks for Leadership

Departed employees achieve virtual immortality, managers scold algorithms—and amidst it all, HR asks how humans and machines can actually be led together. Welcome to the age of great AI confusion.

Helmut patiently answers every question from his colleagues. He knows exactly how the former owner ran the mid-sized company and is familiar with the specific preferences of every customer. As the former head of sales, he willingly shares his wealth of specialized knowledge—around the clock, even when he has just set off to go sledding with his grandchildren. This is because, before retiring, Helmut fed all his expertise into an AI-based avatar. Consequently, the new sales team need not hesitate to bombard their virtual colleague with questions.

That sounds like science fiction, doesn’t it? You got me—this scenario is made up. But preserving corporate knowledge using avatars will soon become reality. Jens Nachtwei, an engineering psychologist at Humboldt University of Berlin and a professor at the University of Applied Management, has been researching the automation of work for twenty years and is intimately familiar with these developments: “We are seeing the emergence of ‘knowledge twins’—chatbots that learn from emails and documentation to replicate a person’s typical decision-making or writing style,” he says. “Of course, this isn’t genuine judgment, but it helps when key knowledge holders leave the company.” According to Nachtwei, other firms go a step further by building so-called process avatars—AI agents that take on specific day-to-day roles, such as onboarding coach or compliance auditor. A brave new world of work? That depends. What matters most in this context is transparency: it must be clear that one is interacting with artificial intelligence rather than a human. Nachtwei also emphasizes that teams must be able to trace the origin of the data and identify who is responsible for it at all times.

Unfortunately, that is not yet the case everywhere when it comes to integrating these smart new helpers. But let’s start at the beginning.

The Great Uncertainty

AI is the driving force behind a massive cultural and organizational shift. Of all the zeitgeists currently shaping our times, this transformation is arguably one of the most powerful. Never before have so many routines, roles, and rules been called into question all at once. Naturally, we cannot simply process this mentally overnight. Even though some may feel they have already figured out the game, technological development is still in its infancy. We find ourselves right in the midst of the era of AI confusion—a fact we should acknowledge first and foremost.

The topic of leadership and responsibility serves as a prime example of this—such as when experienced executives try to treat their AI assistants like human beings. At a US company, a manager recently snapped at an AI tool after it deleted a database, acting exactly as if he were reprimanding a junior employee. The incident has a comical side, too: when asked about the reasons for the blunder, the tool reportedly admitted that the mistake had been a panic reaction. Jens Nachtwei has observed a similar trend: “Some executives are starting to treat AI systems like colleagues. They scold them and expect apologies.” However, this phenomenon is more prevalent in the US, China, and Japan—places where the approach to technology, and the attributes ascribed to AI, differ from those elsewhere. Nachtwei notes: “From a psychological standpoint, this is understandable, but from an organizational perspective, it is tricky. Responsibility must never become blurred; ultimately, a human being must always be held accountable.”

While AI is being humanized on the one hand—perhaps partly because we feel silly talking to machines?—other deeply human tasks are being mechanized. Today, for instance, managers sometimes use AI to analyze and streamline their own conversations with employees. Picture a manager asking a soulless algorithm whether she conveyed her message sensitively enough when discussing a formal warning with a problematic staff member. Isn’t it ironic?

Contradictors rather than sycophants

Ultimately, the responsibility remains with humans. Alicia von Schenk emphasizes this mantra as well. As a junior professor of Applied Microeconomics at the University of Würzburg, she studies how AI transforms decision-making behavior and organizations. She states: “Artificial intelligence is increasingly being integrated into decision-making processes—ranging from HR matters to strategic planning. Yet AI possesses neither social intent nor moral responsibility.” Her research has confirmed that key values ​​such as cooperation, fairness, and trust are contingent upon conditions rooted in human reciprocity—not algorithmic logic. An AI cannot feel empathy or negotiate social norms.

If AI stands by to serve us all day long—praising our brilliant ideas but never contradicting us—this can have further consequences. AI systems learn from our data, mirror our preferences, and thereby unconsciously reinforce our existing thought patterns. In effect, we spend our days fine-tuning a highly personalized echo chamber. “This can diminish cognitive openness and critical reflection within teams,” warns von Schenk. Applied to leadership, this means that if AI always offers agreement, it undermines the social function of criticism and debate—elements crucial for innovation and the ability to learn. Nachtwei is also convinced that this constant stroking of our egos can lead to arrogance and a loss of the ability to handle genuine criticism. Creativity suffers, too: “AI tends to reinforce the average. Unless I consciously break that pattern, the result is a lack of originality.” Nachtwei recommends repeatedly casting AI in the role of a dissenter—for instance, by instructing it to first identify counterarguments or weaknesses. Otherwise, critical thinking is at risk.

Labor market under pressure

But where do we—the confused—actually stand, all things considered? “AI-Based Work Environments 2030” – that is the title of a report that Klaus Burmeister.Which a futurist and founder of the Berlin-based Foresightlab devised a few years ago. That was in 2019, so well over half the journey to 2030 has been covered—time for a progress check. Of the predictions made back then, it is above all the sense of ambivalence that defines today’s reality. “AI is evolving dynamically, yet it runs up against the limits of traditional organizational and leadership structures,” observes Burmeister. Its potential often remains unrealized: while new tools and solutions are hotly debated, blind, reactive activity prevails in many places—things are tried out, piloted, and then discarded. At the same time, surprisingly little is happening elsewhere. Structures are resistant to change, old routines persist stubbornly, and professional training and personnel development seem as though the world got stuck in 2015.

It is that strange state of limbo in which little genuine future emerges amidst the tension between hype and inertia. Burmeister identifies one of the reasons as a clinging to outdated narratives: companies still tell themselves the story of the German economic miracle and the humanization of work—as if transformation were a closed chapter. Yet digitalization and AI call for new narratives and a new culture.

He would like to see HR, in particular, demonstrate a greater sense of realism and the courage to plan ahead. Not every newly launched tool is a miracle cure; what matters is what truly benefits the specific business model. “Executive teams should take a close look at which technologies are genuinely useful for their business model—and which are not. Then, despite the many unknown variables, they should engage in scenario planning and develop a vision for their business model spanning three to five years,” advises Burmeister. For HR departments, this means building skills strategically, managing professional development, and realigning recruitment efforts. Step by step, this creates an organization that does not shy away from uncertainty but instead meets it with a proactive, constructive approach—and one where employees do not primarily associate AI with a threat.

However, the dangers are real. According to a recent warning from none other than US AI researcher Dario Amodei—CEO of Anthropic, the company behind the Claude language model series—AI could replace around half of entry-level office jobs within the next five years. Repetitive tasks in sectors such as technology, finance, law, and consulting are particularly at risk. Amodei criticizes the fact that both companies and governments are downplaying or ignoring the risks associated with AI development. He calls for an open discussion to prepare society for the impending changes. The societal impact should not be underestimated: large-scale job losses give rise to fears regarding social division and economic uncertainty. HR, in particular, must anticipate these changes and develop strategies to support employees, facilitate retraining, and foster new skills.

New Leadership Realities

This state of confusion has long since permeated the day-to-day reality of HR. In recruitment, an AI-driven “arms race” is already underway between applicants and HR departments: cover letters are drafted by language models while models on the corporate side evaluate them—often favoring them over human-written letters. HR expert Nachtwei warns that this leads to a degradation of signals: “Ultimately, applications reveal less and less about actual competencies.” In the short term, this will make the selection process less precise. However, Nachtwei anticipates a shift in the medium term: moving away from cover letters and toward work samples, simulation tasks, and micro-credentials. For HR departments, this means that perfectly polished texts—once a sign that the applicant had invested effort and care—are now of little value. As these facades miraculously become ever more impressive, the challenge is to learn, once again, to look beyond them—to focus on actual performance and competencies, and to gain a personal impression. It is a task that is as demanding as it is rewarding.

Until now, personnel management was an HR function, while system administration fell to IT. Now, these two worlds are converging—hybrid teams comprising both humans and algorithms have become a reality. Alicia von Schenk states: “Leadership must strike a balance between algorithmic efficiency and human judgment. An effective human-AI team is not created through equal treatment, but through a deliberate division of labor and clear responsibilities.” The researcher explains that decision-making under uncertainty typically involves two phases: prediction and judgment—for instance, when considering whether a company should launch a new line of business. While AI systems are becoming increasingly efficient at generating forecasts, the capacity for judgment—encompassing contextual assessment, ethical reflection, and the interpretation of meaning—remains a human prerogative. Moreover, studies indicate that people are more likely to accept algorithmic recommendations if they feel they retain a sense of control. Alicia von Schenk emphasizes: “In the age of AI, leadership therefore does not mean relinquishing control, but rather institutionalizing a degree of critical distance. Openness emerges when dissent—whether from a human or an algorithm—is viewed as a resource rather than a disruption.”

Jens Nachtwei sees entirely new demands facing HR: “I believe we need new ground rules, clear role descriptions within the team, and decision-making protocols. For HR, this means developing hybrid leadership.” In other words, it involves not just leading people, but also facilitating the integration of AI. Essential tools of the trade will include prompt standards and playbooks, as well as an understanding of the systems’ weaknesses. This places high demands on the HR function—requiring it to partially reinvent itself while also keeping a close eye on the constantly evolving tech component. “To be honest, I don’t envy them,” says the researcher.

New formats are also emerging in professional development. Training courses on the critical use of AI, sessions on algorithmic fairness, and workshops on ethical reflection are becoming increasingly important. Many companies are investing in internal academies to equip their workforce for working with machines and algorithms.

The Way Out of Confusion

When will the era of confusion end, and when will the integration phase begin? Jens Nachtwei identifies three conditions: “First, companies need standardized tool stacks—essentially an organizational operating system for AI. Second, they must have clear guidelines—a concise set of AI ground rules that are genuinely put into practice in daily operations. And third, we need AI competencies across all roles.” This process will take years, he notes, but it sets a clear direction for the journey ahead.

Change is unstoppable, yet it can be shaped. HR must become a navigator—steering a course between hype and realism, technology and the human element, science fiction and the near future. Naturally, there is chaos right now—how could it be otherwise? The remedy is to step back from blind, reactive activity, get organized, and devise a plan. When in doubt, seasoned colleagues—who weathered the introduction of the Internet back in the day—might have some advice. Helmut, for one, surely has a wealth of experience to draw upon.

About the Author

Anne Hünninghaus is a journalist and editor at Wortwert. From January to October 2019, she served as acting editor-in-chief of the magazine *Human Resources Manager*. Previously, the cultural and political scientist worked as an editor for the magazines *politik&kommunikation* and *pressesprecher* (now *KOM*).

Sources: humanressourcesmanager.de