What Human-Horse Interactions May Teach Us About Effective Human-AI Interactions

AI systems are progressively making inroads into more application contexts. There is a growing consensus, however, that replacing humans with specialized expertise and implicit knowledge with AI systems may not be beneficial. Some conventional benchmarks for AI, such as the Turing test, might not fully capture the roles AI can play alongside human intelligence since they encourage AI to mimic and implicitly replace human intelligence rather than complement it. Instead of full human emulation by AI, humans should remain “in the loop” in many processes, suggesting that the future of computing depends on a collaborative partnership between human and AI systems rather than a competition. But the question remains: How can this partnership be envisioned, considering the different types of intelligence that humans and AI possess?