In contrast, attributes like explainability and interpretability (49%), and valid and reliable (53%) are viewed with less concern. This says people generally believe the technology works as intended, but their concern is how it behaves.
For the CIO, this means shifting the focus from purely functional metrics to ethical outcomes. A few percentage points of accuracy improvement won’t move the needle on trust. In terms of the privacy attribute, the concern here is profound, especially among end users (69%) compared to providers (53%). This gap requires you articulate to end-users how their privacy is protected not just in general terms, but especially when AI technologies are involved.
2. Target the trust deficit in public-facing scenarios
Trust is not uniformly distributed. The Index reveals that concerns are highest for AI use in media scenarios (339) and personal scenarios (309). Conversely, concern is lowest, and thus trust is highest, in government scenarios (291) and workplace scenarios (289).
