Unpredictable AI

AI technologies are essentially computational systems and as such fairly well understood among IT professionals. However, from social and cultural perspectives, which necessarily include design processes and outputs, these technologies are poorly understood and, as there is no single definition of AI, have become subject to conjecture, speculation and doublespeak. But there are two primary approaches, probabilistic and deterministic AI with each serve unique functions based on their design and outputs: GenAI systems are probabilistic*. Yet AI's long-term impact on, say, education and the labour market is hard to predict because transformative technologies, such as generative AI, take time to become clear. Moreover, there is uncertainty about how humans, both individually and collectively will respond to AI's societal impact, particularly as the AI industry remains largely unregulated creating vast wealth and power inequalities. Given the unpredictable nature of AI, then, designers need to consider both technology and ethics when applying AI systems. *At a high level, probabilistic AI models uncertainty and provides outcomes based on likelihoods. This means that it doesn’t always offer one definitive answer but instead provides a range of possibilities with associated probabilities. Deterministic AI, on the other hand, is rule-based, designed to yield specific, predictable outcomes without room for variability once given a particular input. https://www.dpadvisors.ca/post/the-basics-of-probabilistic-vs-deterministic-ai-what-you-need-to-know