The Artificial Intelligence (AI) paradox

The adoption of artificial intelligence (AI), from routine in every day usage to specialized interventions in dentistry and medicine, from personal needs to sophisticated business applications, has skyrocketed in high-income as well as in developing nations in the last decade. Inequalities remain however, and while AI utilization is practically worldwide, the divide between the northern and southern hemispheres is widening in terms of AI as the foundational infrastructure of modernity, and specifically of contemporary digital life. However, and to some extent paradoxically, the more AI is studied, developed and improved, the more we uncover its weaknesses, limitations, and, as some have argued, its inherent individual and societal dangers. AI generates information based on algorithms that are derived from factual data, which may, or may not have been verified by evidence-based science: information that carries the risk of being biased or fallacious at best, or old and invalidated by new research at worst. Critics of AI also observe its limits, such as in the context of basic human emotional and psychological-social skills. To be clear and as discussed in this writing, the more AI utilization grows and becomes more widespread, the more evident its limitations in reliability and validity become evident.