As Artificial Intelligence systems rapidly transition from passive tools to active teammates in healthcare, workplaces, and education, a critical question emerges: Is it enough for a system to be reliable, or must it also be actively trusted?

A new editorial overview, published in Computers in Human Behavior, explores the complex, multi-layered mechanisms of “trust in technology”—specifically, the alignment between a user’s trust and a system’s actual capabilities. The paper, titled “Interdisciplinary perspectives and current findings on the role of trust as a psychological mediator in human interaction with artificial intelligence: Editorial overview”, was co-authored by Irene Valori, Johannes Kraus, and Merle T. Fairhurst.

The Challenge: Trust is Not a Static Metric

One of the core takeaways of the research is that trust in technology is not a simple, static setting that can be switched on or off. Instead, it is a dynamic, socially constructed process. How much we trust a machine is heavily influenced by:

  • Context: The specific environment and stakes of the situation.

  • Anthropomorphism: The human-like design features or behaviors of the AI.

  • Group Dynamics: How teams interact with and perceive technology collectively.

The Risks of Miscalibration

When there is a mismatch between how much a user trusts a system and how well that system actually performs, serious ethical and functional risks arise.

If trust is too low, valuable automation is rejected. If trust is too high, it leads to dangerous over-reliance—a critical vulnerability in high-risk settings like medical diagnostics, industrial robotics, or classrooms where students might accept AI-generated misinformation without verification.

The CeTI Impact

To maximize the true value of automation, we must design systems not just for peak physical or computational performance, but for the human experience of trustworthiness. Understanding these psychological boundaries is essential for the Cluster of Excellence CeTI as we build the next generation of Human-in-the-Loop systems and cooperative robotics, ensuring that the technology of the Tactile Internet remains safe, intuitive, and properly understood by its users.

The editorial article is available online. Dive into the full overview hier.

Authors: Irene Valori, Johannes Kraus, and Merle T. Fairhurst

This research is published in a special issue of Computers in Human Behavior.