Javad Nouri • Data Scientist • Helvar
During the last few decades, Artificial Intelligence (AI) has helped us perform tasks that seemed nearly impossible just a few years ago. Tasks ranging from detection of cancer to understanding human language. One of the questions in this field is, what is next for AI? Can computers be truly creative? Can AI create art, literature, and invent? We do not have a rigorous definition of creativity, nor do we have a way to measure it, and this is one of the greatest challenges in this field.
There have already been advances in the field of computational creativity in creating art, a discipline that we associate with creativity. Computers can now learn to generate new paintings in a particular style once trained with enough artwork samples from that style. The same goes for generating music, poetry, humour, etc. Deep generative models have helped us achieve these and even mimic works of famous artists.
Scientific inventions is another matter which requires creativity and is more relevant to our topic here. Can one day AI identify a problem, feel the need for some method, device, or process and create it for us?
So far it seems that machines are better at automatic creation of artworks and literature than creativity in science and technology, but if the answer is yes to these questions, it will require us to alter our current processes of dealing with creativity. For instance, our current intellectual property laws will need to extend their horizons to define who owns the rights to artwork, text, or invention created by machines?
Even though computers might not yet be truly creative, they are already being used as tools for inspiring and initiating the creative process in humans. Creativity is still considered to be unique to humans, and if one day we create machines that are creative, their creativity indeed originates from ours.