From a developer's perspective: Psychological safety, ego, and the challenge of transparency in the era of AI
06 | 2025 Arto Kellokoski, Agile Coach & Partner at Kipinä
As an Agile Coach working with various digital development teams, I have learned that a safe environment allows individuals to genuinely embrace vulnerability. But what happens when you hold public authority and responsibility, such as being a CEO? Would you dare to ask an artificial intelligence something you wouldn't otherwise dare to ask anyone? And what about in the presence of others?
We all have an ego designed to protect us, but as a side effect, it is also the most common barrier to individual and organizational growth. It hinders the true potential of artificial intelligence because leveraging AI requires radical self-honesty: What don't I know? Where am I stuck? The quality of your questions determines your insights – and the best questions are usually those you wouldn't dare say out loud.
In these situations, I often find myself returning to the words of actor Eero Milanoff, which stuck with me from an interview in Helsingin Sanomat (May 2, 2021):
"If you cannot mentally open yourself up to your work and the situations you encounter, you are forced to put up defenses. When you are forced to defend yourself, your choices stem from fear rather than the present moment. That cuts off opportunities to create anything interesting."
I believe the same wisdom applies to working with AI: if you are afraid of appearing ignorant, you will never gain access to what you could otherwise know. Instead of putting up defenses, leaders need curiosity and the courage to look at themselves and their decisions in the mirror, including through the help of artificial intelligence.
Whom do you ask, and what?
In my work, I have started building small AI personas – agents, each with its own perspective and context. I use them to challenge my own thinking, introduce multiple viewpoints, and make hidden biases visible.
After all, the answer always depends on whom you ask and how you ask. This is no different from real life. Just like when talking to people, interacting with AI requires understanding what kind of "personality" is responding and the worldview within which it operates. In other words, for an AI agent to truly support decision-making, it always needs context: why this perspective is relevant, and what it brings to the discussion.
As a leader – or even as an Agile Coach – it is important to challenge your own decision-making from multiple angles. Artificial intelligence is a great tool for this. By bringing in multiple agents, such as a critic, a contrarian, an optimist, and an ethicist, you can reflect on your ideas from various directions. Not because any of them is inherently right, but because they help you ask better questions. Sometimes I also have the agents discuss with each other, as if seating them around the same table. For example, when pondering how to deliver difficult feedback, one agent brings up a direct approach, another focuses on learning, and a third asks what might be left unsaid. This kind of small AI workshop often sparks insights I could never reach on my own.
As a leader, it is important to leverage artificial intelligence to practice challenging your own decision-making using these various perspectives. By introducing multiple 'agents' – a contrarian, a critic, an optimist, an ethicist – you can test your thinking and seek a more balanced, deliberate viewpoint. Not because the AI provides the right answer, but because it helps you formulate better questions.
A new layer of interaction
The way people relate to and perceive things directly impacts the outcome. The same applies to artificial intelligence – the interaction we have with it is full of human choices, interpretations, and emotions. In my work with teams, I use AI for iteration and ideation. It helps me structure problems that are often rooted in social dynamics: misunderstandings, tensions, and unclear roles.
I find these challenges particularly prominent in remote work, where much of the interaction takes place through written communication channels. Digital information becomes fragmented, tones of voice and body language are lost, and tacit knowledge rarely surfaces. That is why I have personally integrated AI to help gather the fragments of interaction and synthesize them into a coherent whole—detecting subtle nuances that might otherwise slip by entirely in Slack channels.
When discussing the leverage of AI to enhance interaction, it is crucial to emphasize the importance of self-regulation and awareness. For me personally, AI is not only a tool but a mirror. It does not merely streamline my work; it helps me pause and reflect: Why do I think this way? What if I looked at this differently? This transforms AI into an emotional counterpart—not a replacement for humans, but an extension of my own thinking.
The messy nature of humanity and the crisp logic of machines
To be human is to be inherently messy. Perfection is not the goal—neither in humans nor in AI. Imperfection makes us comprehensible and fosters connection with one another. Presence is a far more important objective than mere efficiency.
Amid the immense potential of artificial intelligence, it is wise to keep its potential for misuse in mind. As efficiency increases, new distortions can emerge within teams, corporations, and human interactions. Alongside the power of AI, the freedom to make mistakes remains one of the most valuable assets in development, and preserving empathy is one of our most critical values. When humans remain at the center of interaction, AI can serve as a support mechanism—not merely as an efficiency booster, but as an enabler of reflection.
I return once more to the words of Eero Milanoff, which capture the core essence of this text:
"When forced to defend oneself, choices stem from fear rather than the present moment. This closes off opportunities to bring something compelling to life."
My personal objective is for artificial intelligence not to build stronger walls and defensiveness, but to help us explore what is compelling, unfinished, complex, and yet meaningful.
Arto Kellokoski, Agile Coach & Partner at Kipinä
Arto Kellokoski is an experienced Agile Coach with a background of over 20 years in software development. He combines technical expertise with agile best practices to help teams and organizations improve their interaction, collaboration, and working methods with the goal of sustainable development. People and genuine interaction are at the center of Arto's work: he listens actively, supports change, and fosters team growth with an empathetic approach.