Post
JA EN

The Era of the Unteachable Leader — The Trap of Mindset Coaching and the Choice of Co-Learning

Posted
By
32 min read
The Era of the Unteachable Leader — The Trap of Mindset Coaching and the Choice of Co-Learning
  • Intended audience: Tech leads and managers who feel they “can’t fully teach the technology” anymore, and junior/mid-career engineers who feel “lately, the feedback is more about attitude than about code”
  • Prerequisites: None
  • Reading time: About 22 minutes

Overview

“Lately, the technical feedback has dried up and it’s mostly about my mindset” — if you’re a junior or mid-career engineer who feels this, it’s probably not paranoia. It’s likely a structural phenomenon. Conversely, if you’re a leader thinking “I can’t fully explain the framework my reports are using anymore,” that isn’t personal decline either.

The junior who expects to be taught and the senior who feels obligated to teach are both bound by the same fixed-role assumption. In reality, depending on the technical domain, the junior often surpasses the senior in domain knowledge. Reverse mentoring — designed around exactly this cross-generational knowledge exchange — has been studied since Jack Welch introduced it at GE in 1999, and the design of juniors teaching seniors has been shown to be effective for organizational learning1. Yet when both parties leave the role assumption unexamined, one carries an “entitlement to be taught” and the other an “obligation to teach,” and the relationship locks them into mutual suffering.

The pace of technological change is structurally outrunning the pace at which leaders can learn. The World Economic Forum’s Future of Jobs Report 2025 projects that 39% of workers’ core skills will change by 2030, and 59% of the global workforce will need reskilling or upskilling2. The skill set that a leader actively used during their hands-on years is becoming obsolete faster than in past generations. Yet organizations still drag the old assumption that “the leader is technically superior to the report.” When leaders try to close this gap-with-reality, their feedback slides off the technical surface and drifts toward “your way of thinking,” “your attitude,” “your mindset”. And mindset coaching, unlike technical coaching, leaves the receiver with no escape. This is exactly where the path to unintended harassment opens up.

The solution is not “leaders catching up.” It is dismantling the fixed teacher/student relationship and redesigning it as a co-learning relationship. Edmondson’s foundational psychological safety study3, the meta-analysis covering over 22,000 people4, and Owens & Hekman’s leader humility research5 all show that “an environment where you can say ‘I don’t know’” and “leaders visibly demonstrating humility” raise team learning and performance.

This article covers how the “unteachable” state is structurally generated, why retreating to mindset coaching tends to convert into harassment, and five concrete behaviors for shifting toward co-learning. Finally, it goes one step deeper than the “look after / be looked after” frame, drawing on Self-Determination Theory’s autonomous motivation to discuss what it means to choose work based on whom you want to engage with.

1. The Structure of the “Unteachable Era”

The premise “the leader should be technically better than the report” was once easy to satisfy. When the half-life of technology was long and five-year-old knowledge was still relevant, it was natural for the older person to lead in volume of expertise.

That world is gone. LLMs, AI agents, vector DBs, new frontend frameworks, the next generation of IaC — first-year engineers run technical stacks that didn’t exist when the leader was hands-on. As Camille Fournier repeatedly emphasizes in The Manager’s Path, management in tech is a separate professional layer that does not require being “the most technically competent” person in the room6. A manager’s job is not to outpace reports in knowledge volume. It is to design the environment in which reports can grow.

Reverse mentoring research points the same direction. Murphy’s 2012 Human Resource Management article frames the design of juniors teaching seniors as a mechanism that bridges the cross-generational knowledge gap and develops both parties’ learning and leadership1. The concept dates back to Jack Welch introducing it at GE in 1999, and it has functioned as an organization development practice for over a quarter century. What matters is that the very fact that this design “works” reflects an embedded premise in OD research: in specific domains, juniors can and do surpass seniors in knowledge.

That said, the experienced person’s advantage doesn’t disappear entirely. Repeated execution automates processing and frees cognitive resources for higher-order judgment. Past investment in deliberate effort underwrites speed. These are real. But Ericsson and colleagues’ classic 1993 Psychological Review paper showed that what produces expertise is not “years of experience” but deliberate practice — engaging with weak points under clear goals, getting immediate feedback, and improving7. The paper has been cited over 9,000 times and forms the foundation of expertise research. A subsequent meta-analysis by Macnamara and colleagues reported that deliberate practice explains less than 1% of variance in professional domains, adding the caveat that practice volume alone doesn’t determine everything8. What is consistently supported across this literature is that “how you spent the time” matters more than “the length of the time”.

So the experienced person’s advantage emerges not from “time itself” but from “the quality of engagement invested in that time.” Experienced people who kept up deliberate practice still hold a clear advantage. But experience without quality engagement does not automatically translate to advantage as years accumulate. And there is no structural guarantee that the older person necessarily has higher “quality of thinking” than the younger one. The automatic linkage of “years of experience → instructor role” wobbles one more notch.

But this redefinition is not always complete in how organizations perceive themselves. Player-manager culture, communication norms that tolerate technical one-upmanship, seniority-based authority structures — in organizations where these elements remain, the old model “leader = technically superior” can be tacitly preserved. As a result, leaders carry pressure to “instruct” in technical domains they realistically cannot keep up with.

Here, the first distortion is born.

2. “Mindset Coaching” as Compensation for Inability to Teach

When the technical domains a leader cannot directly instruct in expand, what’s easy to reach for as material for “instruction” is attitude, posture, mindset.

  • “Take more ownership”
  • “You lack initiative”
  • “Your learning posture isn’t there”
  • “Think with a business perspective”
  • “Your thinking is naive”

Once these become the main axis of instruction, the center of gravity of evaluation shifts from technology to mindset. The problem is that this kind of feedback is often unfalsifiable. Told to “show more initiative,” the receiver doesn’t know what or how much would qualify as “having initiative.” The standard rests in the instructor’s subjectivity, and the evaluation is held permanently below the bar.

Japan’s Ministry of Health, Labour and Welfare published a workplace power-harassment guideline in 2020 that organizes workplace power harassment into six categories. Within “psychological attack,” it specifies as examples “language that denies the target’s character,” “repeatedly delivering excessively long, harsh reprimands related to job execution,” and “repeatedly delivering loud, intimidating reprimands in front of other workers”9. Once mindset coaching evolves into repeated unfalsifiable criticism, the structural distance to falling into this category shortens.

What matters is that the question of the instructor’s subjective intent is not the essence of the problem. In most cases, leaders believe they are giving well-intentioned guidance and don’t even register that they are compensating for their inability to keep up technically. But the receiver’s experience is a continuous rain of irrefutable criticism.

What happens to the receiver in this state has been studied for decades in organizational psychology under a separate research lineage called “abusive supervision.” Tepper’s foundational 2000 Academy of Management Journal paper (cited over 6,900 times) demonstrated that subordinates’ perceptions of sustained hostile verbal and nonverbal supervisor behavior — “ridicules me,” “puts me down in front of others,” “tells me my thoughts or feelings are stupid” — predict declines in job satisfaction and life satisfaction, increases in psychological distress, and a significant rise in turnover intent10. Critically, Tepper’s definition measures “the subordinate’s perception of sustained hostile behavior,” not “the supervisor’s intent.” Intent and structural outcome run on independent tracks.

So the path by which a receiver who’s continuously labeled “below the bar” comes to internalize “I cannot be valued at this organization” is not a leap of inference. It is the empirical structure that a quarter century of abusive supervision research has consistently shown. Mindset coaching, regardless of intent, functions to push people out of the organization — a dynamic this article will revisit in Section 9.

3. Why Mindset Coaching Removes the Escape Route

The decisive difference between technical instruction and mindset instruction lies in whether the claim is falsifiable. With technical instruction, when someone says “this code has an N+1 problem; solve it with preload,” the receiver can engage on the basis of fact. Changing the implementation resolves the comment, or proposing another valid solution moves the discussion forward. The instructor and the receiver stand in the same factual space.

Mindset instruction is different. “You lack ownership,” “your business perspective is shallow,” “I don’t see growth motivation” — these evaluations float in the evaluator’s subjectivity without being made concrete about what would resolve them. The receiver doesn’t see “what to do next” and starts internalizing “what’s wrong with me.” As abusive supervision research has consistently shown, exposure to this kind of repeated denial directed at character or competence predicts elevated psychological distress and turnover intent10.

Schein offered the concept of “Humble Inquiry” against this structure11. Rather than speaking as “the person with answers,” the leader builds the relationship as “the person who genuinely asks questions they don’t know the answer to.” Humble Inquiry is not a “let’s just ask questions” technique. It is a relational design that intentionally flattens the power gradient between instructor and receiver.

When the subject of questioning flips, the dialogue returns to factual space. “Tell me what you were thinking with that implementation.” “I saw it as X, but was there a Y reason?” Questions of this kind contain room for the instructor to update their own understanding. Where mindset evaluation was a one-directional verdict, Humble Inquiry generates bidirectional discovery.

The effect of “the leader visibly demonstrating humility” is empirically confirmed. Owens & Hekman’s Academy of Management Journal study showed across three different research designs (controlled experiment, longitudinal simulation, healthcare field study) that leader humble behaviors “infect” followers, generate collective humility across the team, raise the team’s “collective promotion focus,” and improve team performance5. Leader humility is not an individual personality trait. It is an intervention via observable behaviors that changes the team’s structure.

4. Co-Learning as an Alternative

The solution to the “unteachable era” is not for leaders to chase every technology. That is structurally impossible, and the anxiety of trying to catch up and failing is precisely what accelerates retreat to mindset coaching.

What is needed instead is dismantling the “teacher/student” fixed relationship and redesigning it as a co-learning relationship in which both parties face the same unknown.

Co-learning does not mean leaders “don’t have to learn.” Quite the opposite — leaders take on the role of continuously displaying a learning posture. The difference is that they stop performing as “the person with answers.”

This redesign is supported by the research on psychological safety. Edmondson’s 1999 Administrative Science Quarterly paper, studying 51 manufacturing teams, demonstrated that “the belief that the team is safe for interpersonal risk taking” raises team performance through learning behaviors3. Frazier and colleagues’ 2017 meta-analysis combined 136 independent samples covering over 22,000 people and roughly 5,000 groups, showing that psychological safety has significant effects across diverse outcomes including task performance, organizational citizenship behavior, learning, and commitment4.

In organizations where the leader can declare “I don’t know either,” members surface knowledge and tackle the unknown together. In organizations where the leader keeps performing as “the person who knows,” members hide their ignorance, cover their mistakes, and the organization’s overall learning rate falls. Owens & Hekman’s research shows this gap is interventionable through the leader’s visible humility behaviors5.

Co-learning is a relational design that intentionally embeds these empirical findings.

flowchart TB
    A["Old model<br/>Teacher ↔ Student"]
    B["Premise breaks<br/>Tech change rate<br/>＞ Leader learning rate"]
    C["Retreat to mindset coaching<br/>Unfalsifiable evaluation rises"]
    D["Co-learning model<br/>Peers facing the same unknown"]
    E["Psychological safety<br/>＋ Humble Inquiry"]
    F["Org learning rate rises<br/>Harassment-conversion risk falls"]

    A --> B
    B --> C
    B --> D
    D --> E
    E --> F

5. Five Concrete Behaviors That Support Co-Learning

So that co-learning doesn’t end as fluffy attitude-talk, here are five operable behaviors, each tied to evidence.

(a) Declare “I don’t know either” up front

At the start of meetings or 1:1s, the leader saying “I’m still shallow on this domain” functions as the starting point for psychological safety. Owens & Hekman’s research shows that leader humble behaviors change team-member behavior through social contagion5. Members who watched a leader say “I don’t know” become more able to say “I don’t know” themselves. This is not a confession of personal weakness. It is an intervention that elicits team-level learning behavior.

(b) Make the learning process visible

Deliberately share with the team the leader’s own process of trying new technology — implementations that failed, documents consulted, dialogue logs with ChatGPT or Claude, experimental notebooks. The “learning behaviors” Edmondson 1999 defined include seeking help, asking for feedback, talking about errors, and trying new behaviors3. Showing the in-progress state, not the polished conclusion, normalizes these learning behaviors as “things you can be seen doing” within the organization.

(c) Publicly acknowledge what you learned from a report

The leader repeatedly saying “I learned this from A-san” in public settings is a device for intentionally flattening the knowledge hierarchy. Murphy’s reverse mentoring research shows that designing juniors teaching seniors simultaneously develops the junior mentor’s leadership and the senior mentee’s knowledge update1. A leader who can publicly say “I was taught” formally establishes within the organization the role of “junior who teaches.”

(d) Direct feedback at “process,” not “result”

Rather than “results are insufficient” or “your mindset is insufficient,” focus on “how you thought about that choice” and “what you tried and what you learned.” Wong’s Tripartite Encouragement Model holds that cognitive and affective acknowledgment directed at process — not at character or outcomes — supports self-efficacy and long-term engagement12. This is not “indulgence.” Probing process quality demands more rigorous dialogue than chasing outcomes.

(e) Build settings that flatten the knowledge hierarchy

Reading groups, pair programming, joint research sessions, internal study groups — routinize venues where people learn the same topic without bringing in titles or years of experience. Psychological safety has been consistently measured as a team-level shared belief, not an individual disposition, since Edmondson 199934. Co-learning needs to be embedded in the organization as a system, not as an event. The key is that the leader joins these venues as “a learner,” not as “a teacher.”

6. Risks and Limits of Co-Learning

When discussing co-learning, three misreadings must be avoided.

First, co-learning is not the abandonment of decision responsibility. Learning posture and decision responsibility are separate layers. The leader can say “I don’t know either” while not stepping down from final responsibility for technical selection or directional decisions. Indeed, the ability to make accountable decisions under incomplete information is the core of what is left for the leader. A leader who, under the banner of “co-learning,” lobs “let’s all decide together” is harming the team in a different way.

Second, co-learning is not a license to “not learn.” It’s not that “the unteachable era” excuses leaders from effort — rather, continuously displaying a learning posture is a necessary condition. The moment the leader stops learning, the flatness of co-learning collapses into “the leader does nothing yet retains authority.”

Third, in environments where short-term results are required, applying co-learning across all domains is not realistic. Emergency incident response, releases against tight deadlines, the customer-facing front line — these settings need the structure of quickly delegating decisions to experts. Co-learning is properly run with a designed split: “experts for short-term, co-learning for long-term.”

Without satisfying these conditions, “co-learning” becomes a buzzword and functions as the abandonment of leadership. Co-learning is a relational design in which a person who holds responsibility, while showing a learning posture, faces the unknown alongside others — anything but easy.

7. When You Let Go of “the Leader Who Has Answers”

The norm “the leader should have the answers” remains thick in many organizations — particularly in environments that have tacitly assumed the structure of older people showing answers and younger people receiving them. In such organizations, the leader declaring “I don’t know either” often carries cultural cost. It risks being read as “unreliable” or “lacking leadership.”

But Owens & Hekman’s research shows the opposite outcome. Leader humility, far from lowering team performance, generates collective promotion focus through social contagion and consequently raises performance5. Schein’s Humble Inquiry points the same direction11. Continuing to perform as “the person with answers” is, in technical domains with short half-lives, a fatal liability for the organization. Because that performance intentionally lowers the organization’s learning rate. In an organization where no one can say “I don’t know,” there is no collective exploration of the unknown.

When you let go of “the leader with answers,” another role emerges. The role of designer of the venue where the unknown is faced. Who studies which topic, where learning is shared, what is decided and what is held — instead of having the technical answer, the role is to manage the learning process itself.

This is not easier than “having the answer.” It is harder. More than authority-driven management that pushes through with technical seniority, the quality of design and the quality of relationship building are tested. But this is the only direction in which the management of the “unteachable era” finds an exit.

8. Releasing Role Fixation — Toward Relationships Built on Autonomous Motivation

The discussion so far has centered on redesigning the leader’s role. But a deeper question remains. The “teach / be taught” relationship is often institutionalized in workplaces as a broader role of “look after / be looked after” — an inclusive framework of involvement that includes instruction, evaluation, progress management, listening to worries, and career conversations. The deeper question is whether this “look after / be looked after” framework itself is becoming dysfunctional in the current era.

As reverse mentoring research presupposes, depending on the technical domain, the junior surpasses the senior in knowledge, and learning designs built on this premise function effectively1. A new graduate who knows AI agent design more deeply than a 20-year-tenured department head; a mid-career hire of three years who knows a particular framework’s implementation better than the CTO. The hierarchy of years no longer aligns with the hierarchy of knowledge.

And yet the role expectation “seniors should teach / juniors should be taught” persists as organizational inertia. Seniors carry guilt about not being able to teach; juniors carry frustration about not being taught — both suffering without questioning the original premise.

What is worth referencing here is Deci and Ryan’s Self-Determination Theory (SDT). Gagné & Deci’s 2005 Journal of Organizational Behavior paper organized workplace motivation as a continuum from “extrinsic / controlled motivation” to “intrinsic / autonomous motivation”13. The paper was selected as one of “the eight most influential papers of the past 30 years” in the journal’s 30th anniversary issue and has been cited over 6,000 times. The core finding: autonomous motivation outperforms controlled motivation across engagement, performance, and well-being.

The “look after / be looked after” relationship is established on classic controlled motivation. Seniors engage out of “I must develop the junior” (extrinsic pressure); juniors engage from “I’ll let them teach me” (a passive position lacking autonomy). Neither chooses the relationship — both are placed by organizational structure.

By contrast, “I want to engage” relationships are established on autonomous motivation. “Working with this person is interesting.” “I want to learn from this person.” “I want to face the unknown alongside this person.” From SDT’s perspective, this is the simultaneous satisfaction of the basic psychological needs of relatedness and autonomy — the most healthy expression of intrinsic motivation13.

This implication is not limited to the individual level. When choosing work, the criterion can shift from “is there someone who’ll look after me” or “can I avoid having to look after others” to “is there someone I want to engage with”. When choosing teams, projects, or even when changing jobs — put the desire to engage at the center.

Of course this shift comes with caveats. Not all work can be done “only with people I want to engage with,” and there are organizational situations in which one must work with people one does not want to engage with. But just shifting the axis from “presence of caretaking” to “desire to engage” changes the quality of the relationship significantly. From SDT’s framework, one can reasonably infer the asymmetry: relationships built on obligation (control) tend to elicit evaluative, controlling feedback styles (mindset coaching being one form), while relationships built on autonomous motivation tend to elicit exploratory feedback styles (Humble Inquiry being one form)13.

This also has implications for organizations. The closer the design gets to “employees can work with people they want to engage with on topics they want to engage with,” the higher autonomous motivation rises, and the more co-learning emerges spontaneously13. Conversely, as long as the design enforces “look after / be looked after,” the vicious cycle of unteachable leaders and mindset coaching will not stop.

Not the obligation to teach, not the entitlement to be taught — but engaging because you want to engage. What SDT’s accumulated research shows is that only on this foundation does co-learning function in a healthy and sustainable way.

The contrast between mindset coaching and co-learning discussed so far can be observed at the organizational level as a more general dynamic. How this dynamic separates organizations and individuals in the AI era — that becomes the final question.

9. Two Directions Organizations Move in the AI Era — The Difference Between Exclusion and Support

As organized in the previous section, mindset coaching structurally functions to “push people out of the organization,” while co-learning functions to “keep people in the organization and help them learn.” These two dynamics, beyond individual leader behaviors, form an opposing axis observable as overall organizational culture. “Organizations where those who want to grow are supported by those around them” versus “organizations where dynamics of trying to exclude someone are at work” — which way an organization tilts will substantially shift the competitive outcome in the AI era.

Asserting dominance, sifting out, selecting — when the half-life of technology was long and who held intellectual advantage was relatively fixed, such behaviors had room to be justified as “selection.”

But the era no longer permits this. The World Economic Forum’s Future of Jobs Report 2025 reports that by 2030, 59% of the global workforce will need reskilling or upskilling, and 11% are expected not to receive sufficient opportunities2. Over 120 million workers face medium-term redundancy risk. AI and big data, analytical thinking, and curiosity and lifelong learning are consistently cited as the skill clusters gaining the most importance by 20302. For both organizations and individuals, whether you can sustain a continuously learnable environment is becoming a matter of life and death.

The Cost of Exclusionary Organizations

The concrete damage exclusionary culture inflicts on organizations has been quantified in recent research. Sull, Sull, and Zweig’s 2022 MIT Sloan Management Review analysis used 34 million online employee profiles to identify the strongest predictor of attrition during the Great Resignation. The result: toxic culture is a 10-times stronger predictor of attrition than compensation14. The “Toxic Five” elements they identified are disrespectful, noninclusive, unethical, cutthroat, and abusive14. Read against this article’s theme, unfalsifiable mindset coaching and exclusionary behaviors conceptually overlap with at least three of the Toxic Five — disrespectful, noninclusive, cutthroat — though this is not a causal finding directly demonstrated by Sull et al., but a definitional-level interpretation. The direction is consistent with what abusive supervision research10 has shown about elevated turnover intent.

Exclusionary culture, in other words, is not merely “uncomfortable.” It generates measurable attrition costs for the organization and bleeds out knowledge, experience, and relationships. Combined with the reskilling-opportunity shortfall WEF identifies2, exclusionary organizations carry a double deterioration: “They lose learners while the people who remain don’t get learning opportunities.” It is structurally contradictory for an organization that wants to win the learning-rate competition in the AI era to keep expelling its own learners.

The Long-Term Advantage Supportive Organizations Generate

The reverse-direction evidence is equally clear. Garvin, Edmondson, and Gino’s 2008 Harvard Business Review “learning organization” framework lists as its first building block “a supportive learning environment” — composed concretely of psychological safety, appreciation of differences, openness to new ideas, and time for reflection15. The frameworks discussed in this article — co-learning, Humble Inquiry, autonomous motivation — read as concrete pathways for implementing this “supportive learning environment.”

Inclusive leadership research points the same direction. Carmeli, Reiter-Palmon, and Ziv’s 2010 Creativity Research Journal study, with a sample of 150, demonstrated that leader “openness, accessibility, and availability” raise employee creative work involvement mediated through psychological safety16. “Leaders who don’t exclude” function not merely as an ethically preferable choice, but as an intervention that, via psychological safety, lifts performance and innovation.

Add to this Edmondson’s 1999 original paper3, Frazier et al.’s meta-analysis of over 22,000 people4, and Owens & Hekman’s leader humility research5 — all of which have shown that what wins long-term on performance, learning, and citizenship behavior is not the “sifting” organization but the “supporting” organization. This evidence is robust across industries and cultures.

Leaders Themselves End Up on the Excluded Side

So far the discussion has been about organizations losing reports as a cost. But there is another structural cost. Leaders who can no longer teach and who can only do mindset coaching or exclusion become a net cost the organization is no longer willing to carry, and as a result the leader themselves ends up being pushed out by the organization — that structure.

As a meta-analysis on destructive leadership, Schyns & Schilling’s 2013 Leadership Quarterly synthesis of 57 studies showed that destructive leadership behavior lowers subordinates’ attitudes toward the leader, well-being, and individual performance, while raising turnover intent, resistance to the leader, and counterproductive work behavior17. All of these manifest as direct costs to the organization.

Add to this the leadership derailment literature (the phenomenon where leaders fall off their expected career trajectory), which reports that 30 to 67%, on average 47%, of management positions end in some form of failure or derailment18. The consistently cited primary causes of derailment — poor interpersonal quality, failure to adapt to change, inability to support reports’ growth — overlap almost exactly with this article’s themes.

It is therefore not rational for an organization to retain such leaders long-term. Demotion, reassignment, reduced influence, and ultimately departure from the organization — leaders who continue mindset coaching and exclusion increasingly come to be recognized as “a cost the organization carries.” The choice “exclude because you can’t teach” or “push out via mindset talk” simultaneously slows the organization’s learning rate and functions as a path to making yourself the future excluded one.

This is not a threat. It is a structural fact. The more an organization tries to survive the AI era’s learning-rate competition, the less margin the organization has to keep absorbing destructive leadership.

Structural Risk at the Individual Level

Beyond exclusion risk from the organization, exclusionary behavior is also structurally disadvantageous at the peer level — this is not a direct empirical finding but the structural implication of a fluidifying knowledge hierarchy. The more fluid “who holds intellectual advantage in which domain” becomes, exclusionary behavior is the act of accumulating risk that you’ll receive the same treatment when the positions later flip. The person you asserted dominance over today may be tomorrow’s expert in a domain you don’t know. Reverse mentoring research1 presupposes exactly this fluidity structure: “in specific domains, juniors surpass seniors in knowledge.” In an era where it’s not unusual for “the person who taught” and “the person who was taught” to flip within a year, the behavior of trying to fix the hierarchy structurally raises the probability of suffering disadvantage when the flip happens.

“Those who want to grow are supported by those around them” — the organizational-level implementation of this basic posture is the framework of co-learning, psychological safety, Humble Inquiry, and autonomous motivation discussed throughout this article. Conversely, organizations dominated by “trying to exclude someone” dynamics are structurally disadvantaged in the AI-era competitive environment from every angle examined — Sull et al.’s attrition analysis14, WEF’s reskilling projection2, Edmondson and Garvin et al.’s learning organization research315, Carmeli et al.’s inclusive leadership research16.

This is not moralizing. It is a structural prediction grounded in the consistent direction shown across decades of organizational psychology research.

Summary

To leaders who feel “I can’t teach anymore.” It is not a problem of your individual ability — it is a phenomenon of an era in which the rate of technological change has structurally begun to outrun leaders’ learning rate. Trying to fill that gap with mindset coaching makes it easier to step into the path of unintended harassment. Instead, choosing the “show humility as behavior” intervention pointed to by Owens & Hekman5 and Edmondson’s psychological safety research34 tends to raise team performance.

To junior and mid-career engineers who feel “lately, it’s all about my mindset.” That, too, is a structural phenomenon, and it’s likely not your individual problem. As reverse mentoring research1 presupposes, the hierarchy of knowledge is separate from the hierarchy of years. Denouncing leaders won’t change the situation, but you can present the alternative model of co-learning yourself. “Want to investigate this together as a pair?” “Should we start a reading group?” — moves like these are small steps that change the structure.

And beyond that lies a deeper redesign — replacing “look after / be looked after” with “want to engage”. As Self-Determination Theory13 shows, relationships supported by the desire to engage (autonomous motivation) outperform relationships supported by obligation (controlled motivation) across engagement, performance, and well-being. Shift the axis of choosing work, gradually, from “presence of caretaking” to “desire to engage.”

Acknowledging that you can no longer teach is the paradoxical move that regenerates a teaching culture. When the leader lets go of “the person with answers” and the junior lets go of “the person who is taught,” the organization grows a relationship of facing the unknown together. This is not a confession of weakness. It is one of the strongest learning designs for fast-changing eras, pointed to by nearly half a century of accumulated organizational psychology.

And one final note. Only organizations and individuals capable of the basic posture “those who want to grow are supported by those around them” remain meaningfully in the AI era. Behaviors of excluding, sifting out, asserting dominance — these are residue of an OS optimized for slow-changing eras, behaviors that increasingly raise the probability of the bearer themselves becoming the excluded one. Releasing the “teach / be taught” fixation, choosing relationships by desire to engage, and standing on the side that supports growth — beyond that lies the healthy work-relationship of the AI era.

For more on related themes:

References

References are listed in the order their citation numbers appear in the body.

  1. Reverse Mentoring at Work: Fostering Cross-Generational Learning and Developing Millennial Leaders — Wendy Marcinkus Murphy, Human Resource Management 51(4), 549-573 (2012). Peer-reviewed paper organizing the theoretical foundation and implementation requirements of reverse mentoring. 【Reliability: High】 ↩︎ ↩︎2 ↩︎3 ↩︎4 ↩︎5 ↩︎6

  2. The Future of Jobs Report 2025 — World Economic Forum (2025). Official report integrating projections of labor market change, reskilling demand, and the most rapidly rising skill clusters by 2030. 【Reliability: High】 ↩︎ ↩︎2 ↩︎3 ↩︎4 ↩︎5

  3. Psychological Safety and Learning Behavior in Work Teams — Amy C. Edmondson, Administrative Science Quarterly 44(2), 350-383 (1999). Longitudinal study of 51 manufacturing teams. Academy of Management Outstanding Publication in Organizational Behavior Award (2000). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3 ↩︎4 ↩︎5 ↩︎6 ↩︎7

  4. Psychological Safety: A Meta-Analytic Review and Extension — M. Lance Frazier, Stav Fainshmidt, Ryan L. Klinger, Amir Pezeshkan, Veselina Vracheva, Personnel Psychology 70(1), 113-165 (2017). Meta-analysis of 136 independent samples covering over 22,000 people. 【Reliability: High】 ↩︎ ↩︎2 ↩︎3 ↩︎4 ↩︎5

  5. How Does Leader Humility Influence Team Performance? Exploring the Mechanisms of Contagion and Collective Promotion Focus — Bradley P. Owens, David R. Hekman, Academy of Management Journal 59(3), 1088-1111 (2016). Three studies (controlled experiment, longitudinal simulation, healthcare field study). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3 ↩︎4 ↩︎5 ↩︎6 ↩︎7

  6. The Manager’s Path: A Guide for Tech Leaders Navigating Growth and Change — Camille Fournier, O’Reilly Media (2017). Practical book by an experienced tech-domain manager. 【Reliability: Medium-High】 ↩︎

  7. The Role of Deliberate Practice in the Acquisition of Expert Performance — K. Anders Ericsson, Ralf Th. Krampe, Clemens Tesch-Römer, Psychological Review 100(3), 363-406 (1993). Foundational expertise research paper (cited over 9,000 times). 【Reliability: High】 ↩︎

  8. Deliberate Practice and Performance in Music, Games, Sports, Education, and Professions: A Meta-Analysis — Brooke N. Macnamara, David Z. Hambrick, Frederick L. Oswald, Psychological Science 25(8), 1608-1618 (2014). Meta-analysis bounding the explanatory power of deliberate practice. 【Reliability: High】 ↩︎

  9. Guidelines on Measures Employers Should Take Concerning Issues Arising from Statements and Behaviors Backed by Superior Relationships in the Workplace — Japan Ministry of Health, Labour and Welfare (Notice No. 5, 2020). Official guideline defining the six categories of power harassment and concrete examples of psychological attack. 【Reliability: High】 ↩︎

  10. Consequences of Abusive Supervision — Bennett J. Tepper, Academy of Management Journal 43(2), 178-190 (2000). Foundational paper empirically demonstrating the relationship between abusive supervision (subordinates’ perceptions of sustained hostile behavior) and job satisfaction, psychological distress, and turnover intent. Cited over 6,900 times. 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  11. Humble Inquiry: The Gentle Art of Asking Instead of Telling — Edgar H. Schein, Berrett-Koehler Publishers (2013, 3rd edition 2025). Specialist book by the MIT Sloan emeritus professor. 【Reliability: Medium-High】 ↩︎ ↩︎2

  12. The Psychology of Encouragement: Theory, Research, and Applications — Y. Joel Wong, The Counseling Psychologist 43(2), 178-216 (2015). Theoretical review of the Tripartite Encouragement Model. 【Reliability: High】 ↩︎

  13. Self-Determination Theory and Work Motivation — Marylène Gagné, Edward L. Deci, Journal of Organizational Behavior 26(4), 331-362 (2005). Selected as one of “the eight most influential papers of the past 30 years” in the journal’s 30th anniversary issue. 【Reliability: High】 ↩︎ ↩︎2 ↩︎3 ↩︎4 ↩︎5

  14. Toxic Culture Is Driving the Great Resignation — Donald Sull, Charles Sull, Ben Zweig, MIT Sloan Management Review (2022). Study of attrition predictors based on 34 million employee profiles. Defined the Toxic Five (disrespectful, noninclusive, unethical, cutthroat, abusive). 【Reliability: High】 ↩︎ ↩︎2 ↩︎3

  15. Is Yours a Learning Organization? — David A. Garvin, Amy C. Edmondson, Francesca Gino, Harvard Business Review 86(3), 109-116 (2008). Influential article defining the three building blocks of the learning organization (supportive learning environment, concrete learning processes, leadership that reinforces learning). 【Reliability: Medium-High】 ↩︎ ↩︎2

  16. Inclusive Leadership and Employee Involvement in Creative Tasks in the Workplace: The Mediating Role of Psychological Safety — Abraham Carmeli, Roni Reiter-Palmon, Enbal Ziv, Creativity Research Journal 22(3), 250-260 (2010). Empirical demonstration of the inclusive-leadership → psychological-safety → creative-work-involvement mediation model. 【Reliability: High】 ↩︎ ↩︎2

  17. How bad are the effects of bad leaders? A meta-analysis of destructive leadership and its outcomes — Birgit Schyns, Jan Schilling, The Leadership Quarterly 24(1), 138-158 (2013). Meta-analysis of destructive leadership synthesizing 57 studies. 【Reliability: High】 ↩︎

  18. Leveraging Leadership Development to Pre-Empt Leader Derailments — Physician Leadership Journal (2024). Research review of leadership derailment base rates (30-67%) and primary causes. 【Reliability: Medium-High】 ↩︎

Management
Leadership Psychological Safety Co-learning Harassment
This post is licensed under CC BY 4.0 by the author.