Critical Exploratory Investigation of AI Consumption, AI Perception and AI Literacy Requirements

This article presents the results of a multilayered exploratory investigation on the perception, consumption, and regulation of artificial intelligence-based solutions and the experienced understanding of the technology's development and deployment. This research probes the understanding of concepts such as trustworthy and human-centric AI, the criteria that are used when selecting a certain AI system, and its perceived risks. A multilayered exploratory investigation was conducted, including desk research and qualitative input collection. An AI literacy deconstruction exercise was conducted, mapping the literacy needs and the risks associated with low AI literacy levels for three roles interacting with an AI system-developer, deployer and affected person/user, in response to the AI literacy obligations of the European Artificial Intelligence Act. AI literacy programmes must be role specific and calibrated to function and organisational needs. Transversal skills and knowledge are found in all roles, such as critical thinking, needed to correctly interpret the results. The affected citizen/ user also must have a wide range of skills and knowledge of how AI systems function, what their limitations are, the legislation addressing AI systems and ethical values and principles that should be embedded in these systems. It is not low AI literacy in absolute terms that is risky, but rather unfounded perception of one's own high AI literacy and overreliance on AI systems in the absence of critical thinking. AI systems are perceived in a dystopian manner due to their risks and the attention inflation they receive, at the expense of other topics.