Psychological Safety and 3 Levels of Accountability: Part II

Psychological safety without accountability is like a soccer referee without a rulebook: the game still “continues,” but nobody knows what counts as a foul, and outcomes drift toward whoever can push hardest without getting called.

That is the trap some teams fall into when they treat psychological safety as consequence-free expression. People can “speak up,” but there is no shared procedure for calling out coercive delivery, careless generalisations, or subtle put-downs. Over time, the safest strategy becomes silence, sarcasm, or political manoeuvring.

The standard, widely cited definition (Amy Edmondson) is: team psychological safety is “a shared belief that the team is safe for interpersonal risk taking.”

That definition explains this predicament because “safe” is about freedom from humiliation, rejection, or punishment for speaking up, not freedom from being held accountable for the impact of one’s behavior. When teams (or leaders) compress the idea into “everyone can say anything,” psychological safety gets mistaken for comfort, permission, or consequence-free expression.

In practice, the mismatch shows up like this:

• Psychological safety protects the act of speaking up (questions, concerns, dissent, admitting mistakes).
• It does not remove the need for standards about how people speak (tone, coercion, dogmatic rhetoric, disrespect, manipulation).
• Without an agreed accountability mechanism, the loudest or most forceful delivery can dominate, and quieter members learn that speaking up is “allowed” in theory but costly in reality.
• Edmondson has repeatedly noted the concept is often misunderstood; it is compatible with discomfort, candid disagreement, and high standards, rather than being “niceness” or low urgency.

So the predicament is built into the phrasing: “safe for interpersonal risk” can be misread as “safe from interpersonal correction.” The fix is to pair the definition with an explicit norm: speak up is protected; delivery is accountable.

Accountability fixes this, not by policing opinions, but by setting standards for behaviour while opinions collide. SpatzAI frames accountability in three practical levels, applied proportionately and in real time.

Level 0. is a verbal caution: a simple, immediate signal that a line may have been crossed, requesting an acknowledgment. Most issues end here when intent is good and awareness is high. This is akin to the ref offering a player a verbal warning.

Level 1. is the formal caution (SpatChat App): informing the infringing person that the objector will send them a spat at a convenient time, requesting an acknowledgment via the app for documentation. Here too, most issues will be resolved. This is akin to the ref blowing his whistle as an official caution.

Level 2 is a formal objection (SpatChat App): used when the behavior is denied or challenged, requesting a simple apology for the delivery, not a retraction of the underlying view. This is akin to the ref using a yellow card.

Level 3 is a stop and review (SpatChat App and Team + AI Review): when the moment cannot be resolved quickly, it moves to a defined team-and-AI review process so “reasonableness” is determined by agreed method, not by status, confidence, or volume. This is akin to the ref using a red card and sending off the player.

Psychological safety is the permission to play. Accountability is the rulebook that makes play fair.