Leaders who are in danger of losing their position are more likely to take advice from a data algorithm than another human, research from BI Norwegian Business School reveals.

Psychologist Ingvild Müller Seljeseth conducted studies where participants were assigned to either stable or unstable leadership positions, which they would lose if they made a wrong decision. When asked to estimate the number of peas in a jar, leaders in a stable position were far more inclined to accept advice from previous participants than leaders in an unstable position.

Participants were also asked to make estimates of the anticipated price development in the stock market, before being given advice by a data algorithm or previous participant. The threatened leaders were more likely to accept advice from the data algorithm than another human being.

Ms Seljeseth says: “Leaders in an insecure position feel stressed so become rigid in their thinking and stick to their own decisions. They also feel at risk of being perceived as incapable or indecisive if they take advice from someone else. However, these insecure leaders are more willing to accept advice from the data algorithm as machines are not perceived as competing with them, unlike a human being.

“The big concern is that, as leaders have to constantly make decisions that are in the best interest of their organisations, better decisions are made by leaders who are open to advice and suggestions from others.”

The researchers also explored the impact of the reliability of the advice. When it was revealed whether the advice came from a prominent expert or someone of average competence, the leaders in a stable position paid more attention to the expert advice, while leaders in a threatened position remained reluctant to follow advice regardless of whether it was from an expert or not.

The findings indicate that an insecure leader feels too threatened to accept advice, suggesting that those who would gain the most from taking the advice of others, are less prepared to do so.

The findings come from three studies involving a collective sample of around 500 participants.