What can machines teach us in our journey of reproducing human scientific creativity?
Keywords: scientific creativity
TL;DR: We explore the potential for building AI systems that surpass human creativity by leveraging their unique strengths while reflecting on human traits worth emulating, ultimately focusing on what we should create
Abstract: In the race toward creating a strong AI, we have historically focused on replicating human intelligence. For many advanced tasks such as language and image generation, complex classifications in fields such as medicine, computer vision and other sensor data in self-driving cars, we have been successful. However, for complex behaviours like creativity, we often deem machines incapable. Maybe we are desperate to have something of our own, that machines could never do. Maybe we are too prideful in our own intelligence. What if we were tasked to build a truly creative AI capable of intuition and insight? What should we consider? Would replicating human abilities be the best option, or could we make something even better?
This article holds a mirror up to us and explores scientific creativity. We first explore the many properties that may allow machines to surpass humans in creative insight, such as unbounded effort and lack of competition. We should exploit these, rather than limit them in the attempt to make AI more human-like. In the second half of this article, we realise there are many traits we have overlooked in ourselves, that we should strive to emulate in machines. There is no doubt that machines someday could mimic human creativity. The purpose of this reflection is to realise it is not about what we can build, but what we should build.
Submission Number: 4
Loading