By Alice Rickert, Research Associate at the University of St. Gallen
This blog is produced as a part of the Reshaping Work Fellowship Programme. The opinions and views expressed in this publication are those of the author. They do not purport to reflect the opinions or views of Reshaping Work or organisations that have supported the programme.
Every few months, a new HR technology promises to revolutionize leadership: algorithms that save time, cut costs, and either outperform or enhance leaders’ decisions (so-called leadership augmentation). Yet, the approach of minimizing costs through technology usage with equal regularity creates scandals. The Artificial Intelligence Index Report (2025) documents a 56.4% increase in AI-related incidents in 2024 compared to 2023, indicating that rapid AI adoption outpaces responsible oversight and deployment.
The problem is not lack of awareness or intent; most leaders genuinely aim to use AI responsibly. The issue seems to lie in a fundamental misunderstanding of what responsible augmented leadership requires. Organizations often treat it as a matter of awareness or abstract values, overlooking that it also requires the capability to apply those values, balancing competing stakeholder needs and most importantly take the context into account. To understand what responsible augmented leadership needs, we first have to establish what augmented leadership is to then delineate what it requires.
What is Augmented Leadership — and What It Is Not
Leadership augmentation refers to the use of technology to enhance a leader’s decision-making, either by providing advice, concrete recommendations, or partially automating leadership functions. It goes beyond descriptive data presentation, characteristic of conventional leadership, to technology that actively recommends and shapes a course of action
Algorithmic management, by contrast, involves outsourcing leadership decisions entirely to algorithms, often with minimal human oversight. This application of technology is common for administrative management tasks such as shift schedules, route assignments, and other gig-economy contexts. For example, Uber drivers are managed primarily by algorithms.
Organizations are interested in leadership augmentation, as research suggests that successful human-AI collaboration can lead to above-average decision-making outcomes [1].
The High Cost of Augmented Leadership Failure
Consider the stark consequences for the affected individuals. When augmented leadership fails, it often upends an individual’s life fundamentally:
- The UK Post Office Horizon scandal led to the wrongful criminal prosecution of 900 sub-postmasters due to accounting software errors that were ignored by leaders for over 15 years [2].
- Australia’s Robodebt scandal that affected 440.000 welfare recipients intended to save $1.76bn, causing financial devastation, and, in some cases, severe personal distress [3].
- Finally, despite Amazon publicly retiring its biased recruitment algorithm, other organizations faced similar allegations (e.g., Workday, HireVue) [4,5,6].
These cases illustrate the necessity of augmented leadership and the difficulty of maintaining it. The responsible deployment of such technologies did not principally fail due to a lack of awareness. Signs of technological errors have been present across all organizations for years. Rather, they often were failures of leadership training and culture that could have enabled timely intervention (Post Office and Robodebt Scandal) and cross-organizational learning (Biases in Recruitment).
Why Augmented Leadership Fails: Two Inconvenient Truths
To actually lead better with AI, we have to acknowledge two inconvenient truths:
1. It’s not simply “evil leaders.”
My research [7] and industry reports show that most leaders are aware of AI-related risks, HR analytics teams work diligently to mitigate problems, and organizations genuinely intend to act responsibly. Just “bad” leadership exists, but is rarely the main driver of large-scale failures. Critical incidents often emerge despite a genuine intent to use AI responsibly, indicating other structural factors are at play, such as resources, time, leadership training, decision-making processes, or the organizational culture.
2. Mistakes are inevitable.
When organizations experiment with technology, errors will occur. The key question is whether organizations have created safety nets that allow leaders to identify and correct mistakes. This requires trained individuals who can recognize problems and intervene effectively.
Reducing algorithmic errors (e.g., bias) is largely a matter of testing and resources — a solvable challenge. What algorithms cannot do is fully understand context. This turns augmented leadership into a challenge.
What Responsible Augmented Leadership Actually Requires
Responsible augmented leadership demands organizational systems that empower leaders to exercise judgment and intervene when necessary. Our research shows [7] that in practice, some leaders lack either the authority or the capability to intervene appropriately when technology-driven decisions require human oversight. As a result, established organizational processes often undermine trust by optimizing for efficiency and standardizing to the extent that they inhibit the behaviors that augmented leadership requires — such as double-checking, investing time, and critically reflecting on the right course of action. Therefore, leaders face a difficult balancing act:
- They must rely on algorithms appropriately, neither over- nor under-relying.
- They must have the courage to contradict recommendations, even when algorithms are correct the majority of the time. If the systems were to be severely flawed, we would not use them. This makes interventions feel riskier for leaders — what if their intervention is wrong?
- They have to handle ambiguous situations that are rarely clear-cut. Employees may not disclose personal circumstances, and leaders may misjudge the appropriate intervention.
In conclusion, contextual judgment remains essential. Organizations must deliberately design processes and cultures that enable — not just permit — leaders to assume responsibility when using leadership technologies.
Let’s Start Today
To summarize, scandals involving leadership technologies mostly do not emerge from malicious intent but from the absence of appropriate conditions. Even well-meaning leaders systematically fail to translate awareness into responsible action if they are not trained or restricted by their work context. Starting points for yourself or your organization can be the critical assessment the following aspects (see also Table 1 for concrete actions): Do we possess….
- … trained leaders equipped to recognize errors and assume responsibility.
- … clear decision processes that allow intervention and contextual judgment.
- … an organizational culture that encourages critical engagement with algorithms.
Without these, intention alone cannot prevent harm, no matter how aware or responsible leaders aim to be.
