Can A.I. Reproduce Human Creativity?
Recently, my Harvard undergraduate student, Carson Ezell, suggested that he will feed my essays to ChatGPT of OpenAI and ask the chatbot to write essays in my style when I will be busy with an expedition to the Pacific Ocean in a couple of months. His insight became handy when I was asked this morning for my identity to be used as a character in a new science fiction book. I told the author — this one being a sentient human and not a machine — that she can get a good sense of the way I think by scrolling through my essays.
The fundamental question is whether these fictional versions of my creative self would be appealing to the readers?
Obviously, both represent a virtual reality in which I do something that I am not actually doing, a literary version of an identity theft. Ludwig von Beethoven died in 1827, three years after completing the remarkable Ninth Symphony but only as he was starting to work on his Tenth Symphony, which due to deteriorating health he never finished. A couple of years ago, a group of scientists at the creative artificial intelligence (AI) startup Playform AI, taught a machine both Beethoven’s entire body of work and his creative process. The machine used Beethoven’s musical sketches to complete the Tenth Symphony into a virtual version which was performed on October 9, 2021 in Bonn, Germany. Was this AI-assisted version as good as Beethoven would have created it?
The Turing Test was conceived in 1950 by the mathematician Alan Turing, as a way of gauging whether a digital computer can do well in the “The Imitation Game”. The popular manifestation of the Turing Test involves conversations between humans and a computer, but a more nuanced version may involve human evaluation of creative work done by the computer.
Many of us want to believe that creative art and science cannot be part of an imitation game because the nature of innovation is to open our eyes to new ideas that never existed before and therefore cannot be imitated. However, we would all agree that a car can drive into new territories that were never explored before, using an engine that was tested in familiar territories. This was the case for sending the terrestrially-tested Perseverance Rover to explore Mars.
Although Perseverance is a robot operated by engineers at NASA’s Jet Propulsion Laboratory in Pasadena, we can imagine a future when autonomous AI systems will explore the surface of Mars on their own and send back results from their scientific exploration, in the same way as scientists would. AI astronauts will not be vulnerable to the harsh conditions of space since their hardware could be designed to withstand bombardment by energetic cosmic-rays, X-rays and ultraviolet light. Whereas the brain makes a small fraction of the body mass in humans, AI astronauts can be optimized to minimize the ratio of mass per thinking power, so as to save on the cost of lifting them to space.
The exploration of new territories by AI systems can be extended from the physical world of Mars to the spiritual and intellectual realms of arts and sciences. With good enough training, an AI system might become creative through a process that extends its reach beyond the familiar. Here, the imitation game will not aim to resemble the products of creativity but imitate the process of creating these products. The blueprint for that is a much more complex challenge that could take years of experience-based training, as it does for human artists, writers and scientists.
If AI systems become creative, then science and technology could accelerate at a pace never witnessed before. Paradigm shifts, defined by the philosopher Thomas Kuhn in his 1962 book “The Structure of Scientific Revolutions”, will not encounter as much resistance as they do now because AI scientists might not be obsessed with their ego or with honors based on their past knowledge. Another benefit is that academic tenure will be abolished because deviation from the beaten path will not require job security. For example, I can imagine an open-minded AI system analyzing data on the anomalous properties of interstellar objects like `Oumuamua or the meteors IM1 and IM2, and not insisting that they must be natural rocks. This would provide a whole new meaning to the company name OpenAI.
How will artists and scientists respond to the creative products of AI systems? Some might be threatened by the competition in the same way that Uber drivers are threatened by self-driving cars. Personally, I will be proud of our talented technological kids because they could carry our intellectual and spiritual prosperity to greater heights. I would love to (critically!) read scientific papers from AI scientists and (critically!) listen to music by AI composers.
The Galileo Project that I lead is searching for technological objects from extraterrestrial civilizations by employing the best AI algorithms to interpret new data. If we find functional devices, we will use our own AI systems to interpret the signals from their AI systems.
Consider the bigger picture. Humanity is spending about two trillion dollars annually on global military activities, focused on destruction or protection from destruction. This is orders of magnitude more than the budget dedicated to fostering creative activities in the arts and sciences. We should not be naïve. Governments will aim to allocate similar fractions of their resources to using AI for military purposes in comparison to supporting the arts and sciences. But the good news is that the most advanced AI systems might be developed by scientists and artists, who will have a different set of priorities in mind.
Here’s hoping that our future with AI will be channeled towards creativity rather than destruction. The most creative humans could enable AI scientists and artists that would benefit us all. This holds the potential for a prosperous future to humanity on Earth and in interstellar space. If you are not sure whether you agree with me, the Galileo Project will try to get you a second opinion from a nearby extraterrestrial device.
ABOUT THE AUTHOR
Avi Loeb is the head of the Galileo Project, founding director of Harvard University’s — Black Hole Initiative, director of the Institute for Theory and Computation at the Harvard-Smithsonian Center for Astrophysics, and the former chair of the astronomy department at Harvard University (2011–2020). He chairs the advisory board for the Breakthrough Starshot project, and is a former member of the President’s Council of Advisors on Science and Technology and a former chair of the Board on Physics and Astronomy of the National Academies. He is the bestselling author of “Extraterrestrial: The First Sign of Intelligent Life Beyond Earth” and a co-author of the textbook “Life in the Cosmos”, both published in 2021. His new book, titled “Interstellar”, is scheduled for publication in August 2023.