Artificial intelligence as a positive and negative factor in global risk

Abstract

The most significant danger posed by Artificial Intelligence (AI) is the widespread misconception that we fully understand it. This misperception stems from AI’s history of grand promises followed by underwhelming results. Unlike other complex fields like stellar astronomy, AI suffers from an exaggerated sense of comprehension, leading to unrealistic expectations and potential misjudgments. The difficulty of AI lies not in its inherent complexity, but in our tendency to overestimate our knowledge of it. Given the lack of empirical data to quantify AI risks, our understanding is further hampered by cognitive biases, which are amplified under pressure and uncertainty. Consequently, we must be wary of prematurely assuming we grasp AI’s potential consequences, as this could lead to overlooking crucial risks and failing to adequately prepare for the future.