AI adoption needs more than tech: Darya Mishyna on psychological safety

In a recent article published in Learning Magazine, Matthew Syed Consulting’s Principal Business Psychologist, Darya Mishyna, explores why the biggest hurdle to artificial intelligence integration is human behaviour – specifically team dynamics.

While companies invest heavily in AI infrastructure, technology alone cannot evaluate complex trade-offs between speed, accuracy, risk, and ethics. To unlock the full value of AI, organisations must foster psychological safety alongside diversity of thought.

The 4 team dynamics shaping AI implementation

Drawing on examples from Pixar’s famous “Braintrust” to global tech innovators, Darya outlines four distinct team dynamics that emerge when organisations attempt to implement AI. Understanding where your teams fall is key to mitigating risk and accelerating performance.

1. Creative Friction (the ideal state)

Diverse cross-functional teams (combining data science, legal, HR, and operations) openly debate assumptions and challenge trade-offs. This high-safety, high-diversity environment surfaces hidden risks early and yields responsible, high-impact AI solutions.

2. Stagnation

Teams lack both diversity of thought and the safety to speak up. Decision-making feels smooth, but decisions suffer from severe groupthink, leading to poor adoption and unaddressed risks.

3. Untapped Potential

Cognitive diversity exists, but team members hesitate to voice dissenting views or non-technical concerns due to low psychological safety. Valuable insights remain hidden.

4. Overconfidence Trap

High energy and active debate mask a lack of perspective diversity. For instance, a team composed solely of software engineers may rigorously test model performance while completely missing operational, legal, or end-user risks.