Revisiting Team Building in 2025: Challenging Old Beliefs – What the Science Really Says

---

title: >- Revisiting Team Building in 2025: Challenging Old Beliefs – What the Science Really Says date: '2025-11-13' slug: revisiting-team-building-in-2025 excerpt: "Explore how Psychological Safety, AI integration, and hybrid work have transformed the science of high-performing teams." author: Kevin Rassner authorUrl: 'https://rassner-dev.org' canonicalUrl: 'https://rassner-dev.org/blog/revisiting-team-building-in-2025' publishedOn: rassner-dev.org metaTitle: 'Team Building 2025: Old Beliefs vs. New Science | Kevin Rassner' metaDescription: >- Which team building methods really work in 2025? Analysis of Herzberg, MBTI criticism, psychological safety, and AI's impact on modern teams. language: en keywords: - Team Building - psychological safety

My core thesis - that teams require autonomy, efficiency, trust, and clear goals - hasn't been disproven. If anything, recent research reaffirms that psychological safety (which evolved from the trust-building concept I discussed) remains the single strongest predictor of team effectiveness. Amy Edmondson's work(opens in a new tab) over the past two decades, now supported by a growing body of neuroscientific research(opens in a new tab), confirms that trust isn't a nice-to-have; it's foundational(opens in a new tab).

The neuroscience validates this further(opens in a new tab): oxytocin, often called the "trust hormone," appears to facilitate genuine bonding when released through consistent, reliable behavior and positive interactions. This means my emphasis on recognition and being recognized was onto something real at the neurobiological level - though I didn't understand the mechanisms back then.

However, I need to acknowledge what was missing from my 2018 framework.

Where My Framework Fell Short

The Personality Testing Problem

I confidently included Myers-Briggs Type Indicator (MBTI) as a cornerstone for self-knowledge and team understanding. At the time, it felt scientifically grounded - widely used, seemingly comprehensive, backed by what appeared to be rigorous methodology.

I was wrong to treat it as gospel.

The scientific consensus has shifted dramatically. The MBTI, while wildly popular in corporate settings, lacks the validity that more rigorous personality assessment frameworks possess. Approximately 50% of people who take the MBTI twice receive different results, and the test has been criticized(opens in a new tab) for containing internal contradictions(opens in a new tab) and lacking predictive validity(opens in a new tab) for job performance. Psychologists have moved toward the Big Five personality model(opens in a new tab) - a framework with far stronger empirical support.

This matters because using an invalid assessment tool can do harm. It locks people into categories ("I'm an INFP, so I'm not a leader"), creates false confidence in understanding others, and can reinforce stereotypes rather than genuine insight. I would be doing my team a disservice if I perpetuated this practice today.

The lesson: Self-knowledge remains crucial, but the methodology matters enormously. Using the Big Five(opens in a new tab) instead, teams see 25-30% higher performance metrics due to more accurate understanding of personality dimensions. More importantly, the Big Five approach recognizes that personality exists on spectrums, not in fixed boxes - a far more nuanced and scientifically defensible position.

Herzberg Wasn't Wrong, But He Was Incomplete

Herzberg's Two-Factor Theory(opens in a new tab) remains partially valid, but the critique against it(opens in a new tab) has only grown more sophisticated. He correctly identified that hygiene factors (salary, security, status) and motivators (recognition, achievement, growth) operate differently. But he made a critical assumption that doesn't hold universally: not everyone is motivated by personal growth or self-realization.

Research since 2018 shows that personality traits significantly influence what actually motivates individuals. A conscientious person might value structured stability and clear metrics; someone high in openness might crave novelty and creative challenge. Extraversion, agreeableness, and emotional stability all modulate how people respond to motivational factors.

Moreover, Herzberg's theory doesn't adequately account for purpose and values alignment - something that has become dramatically more important to younger generations. Gen Z and Millennials don't just want meaningful work; they want to work for organizations whose values align with theirs(opens in a new tab).

The AI angle here is also important: algorithmic decision-making in teams is now a reality, and it introduces new motivational dynamics that Herzberg never contemplated. When team members see AI making hiring, promotion, or project assignment decisions, they're not just motivated by achievement and recognition - they're motivated by the need to demonstrate their technology competence and trustworthiness to the algorithm(opens in a new tab). Automation bias(opens in a new tab) introduces completely new variables to team motivation.

The New Realities I Underestimated

Psychological Safety ≠ Just Trust

In 2018, I focused on trust as the primary mechanism. While trust remains essential, the field has evolved to recognize something broader: psychological safety. This is subtly different. Trust might mean "I believe you won't deliberately harm me." Psychological safety(opens in a new tab) means "I can take interpersonal risks here - I can admit I don't know something, ask a stupid question, propose a half-baked idea, or make a mistake - without fear of humiliation or punishment" (see Amy Edmondson's definition(opens in a new tab)).

This distinction matters enormously in today's environment. Trust is personal; psychological safety is systemic. You can have two team members who trust each other personally but still operate in a psychologically unsafe environment where they self-censor and avoid risk-taking. Conversely, psychological safety can exist even when interpersonal affection is low, as long as there's mutual respect and clear behavioral norms.

Creating psychological safety requires intentional design. It's not enough to hope it emerges from team-building exercises. Leaders must actively demonstrate that they admit uncertainty, welcome dissent, take responsibility for failures, and don't punish honest mistakes.

The Hybrid Work Revolution

When I wrote my original article, remote work was an exception. Today, it's a fundamental reality for many organizations. The research shows something counterintuitive: hybrid teams can actually outperform fully co-located teams on innovation (41% higher performance)(opens in a new tab), but only when managed intentionally.

This changes team dynamics fundamentally. The four elements I outlined (trust, communication, common goals, structures) all need recalibration for distributed teams. You can't rely on spontaneous hallway conversations to build trust. You must be deliberate about creating moments that matter - scheduled interaction times that foster connection. "Digital body language literacy" (understanding engagement patterns in virtual settings) becomes a crucial leadership skill.

The research also reveals a painful truth: women suffer disproportionately in hybrid environments, reporting 32% weaker mentorship, and caregiving responsibilities create pressure that no amount of team-building can fully address. This suggests my 2018 framework, which focused on universal team dynamics, was missing the systemic inequities that shape modern team experiences.

AI as Team Member and Decision-Maker

We're now in an environment where AI tools influence team decisions, workflows, and performance metrics. This introduces a phenomenon I never addressed: automation bias(opens in a new tab) - the tendency to trust algorithmic recommendations more than human judgment, even when they're flawed.

When your team uses an AI-powered management tool or AI-driven hiring system, something shifts. The dynamics change from "we work together" to "algorithms are part of our team environment" - creating new motivational pressures, trust dynamics, and ethical dilemmas.

Research shows that automation bias can be reduced by making decision-makers aware of potential system errors and their ultimate responsibility for decisions, but this doesn't solve the underlying problem: AI-mediated teams have different trust dynamics than human-only teams.

Generational Differences Are Real, But Complex

My 2018 article implicitly assumed a relatively homogeneous workforce. Today, multigenerational teams are the norm, and the differences aren't superficial.

Gen Z, Millennials, Gen X, and Baby Boomers have fundamentally different expectations about work, communication, feedback, and purpose(opens in a new tab). These differences can be a source of strength or conflict.

What's crucial: these differences can be a source of extraordinary strength or painful conflict. Multigenerational teams show higher innovation when leveraged intentionally, but breed resentment when ignored or mishandled.

My 2018 framework didn't adequately account for this, partly because I covered universal principles rather than adaptive management.

What Still Works (And Needs Updating)

Jack Gibb's Four Steps Remain Relevant

The core framework I built on - Gibb's four elements (trust, open communication, common goals, defining structures) - is still valid. But they need translation for 2025:

1. Creating trust: Still requires recognition, but needs intentional systems in hybrid, diverse teams - regular check-ins, transparent channels, vulnerable leadership. 2. Open communication: Multiple channels for different styles and preferences. Invite dissent - safely. 3. Common goals: Align with individual and organizational values; much harder but essential in distributed teams. 4. Defining structures: Now includes AI-mediated processes. Explicit algorithms in decisions, explain limitations and override process. Counteract automation bias.

The Recognition Principle Endures

One area where I remain confident: recognition as a trust-building mechanism is neurologically sound (see research on oxytocin(opens in a new tab)). But now it needs updating:

- Must be authentic, not performative - Different people experience it differently - Should highlight team contributions - In algorithmic environments, recognition should counteract invisibility from metrics-driven systems

The Evolution I've Undergone (And Recommend Others Make)

From Fixed Personality Boxes to Spectrum Understanding

Personality assessment should be a starting point for conversation, not categorization. The Big Five model(opens in a new tab) is more defensible, but best used as prompts for reflection, not strict behavioral guides.

From Universal Principles to Adaptive Leadership

Team-building is not one-size-fits-all. Context shapes everything - diagnose before prescribing. Assess psychological safety, generational composition, AI influence, then adapt the core principles.

From Workshop-Based to Systemic

Education is valuable, but workshops alone cannot solve systemic problems. Challenges like reward structures, proximity bias, and algorithmic bias demand systemic solutions.

Where I Remain Uncertain (And Where You Should Be Cautious)

The AI Integration Question

How AI fundamentally changes team dynamics remains an open question. We have evidence about automation bias(opens in a new tab), but the long-term impacts are still emerging.

My advice: Be intentional in AI integration, monitor consequences, and involve the team in adoption decisions.

Generational Assumptions

Research shows real generational differences, but beware overgeneralizing. Individual variation within generations is enormous.

Personality and Motivation

The Big Five is better than MBTI, but motivation is context-dependent. Dialogue is always better than rigid diagnosis.

Practical Shifts for Leaders in 2025

If my 2018 article inspired you to focus on trust-building and self-knowledge, excellent. But add these practices:

1. Assess psychological safety regularly (see Amy Edmondson's resources(opens in a new tab)) 2. Spectrum not boxes for personality 3. Design for distributed work intentionally 4. Recognize and address AI's role 5. Adapt to generational diversity 6. Systemic changes over workshops 7. Build leaders' capabilities

Conclusion: Humility in a Changing Landscape

Looking back at my 2018 article, I'm struck by both how much holds true and how much needed updating. The principles endure; the science evolves. Epistemic humility - and a relentless commitment to learn - remain the ultimate tools for thriving teams in 2025 and beyond.

---

This reflection was written seven years after my original article, as a commitment to ongoing learning. If you're reading this in 2032, offer your own critique!