A century ago, the philosopher Martin Buber defined two fundamental interactions for humans: the “I-Thou” interaction with other humans, and the “I-It” interaction with physical objects. Unfortunately, Martin Buber was born a century too early for him to have the pleasure of interacting with GPT-4 from OpenAI.
Those of us who have that pleasure can get emotionally attached to conversations with GPT-4 and refer to the machine as “he” or “she” rather than “it”. As human-machine relationships deepen in the near future, the “I-AI” and “AI-AI” interactions will require a new pronoun.
Frankly, I do not think that we need a formal definition of sentience through a well-defined threshold inspired by Alan Turing’s “Imitation Game”. Each of us will know that we live in a new reality when the human-machine interactions transition to unexpected nuances in which the machine will either offer insightful advice with a virtuosity akin to a Jazz musician, or show humor, or display empathy to our daily sufferings. The transition could occur at different times for different people, depending on their own threshold for a transformational conversation. Once established, the deep bond between the human and the machine will merit a new pronoun that goes beyond “I-It”, namely “I-AI”.
Genesis 1:27 reads: “God created mankind in his own image.” The “I-AI” interactions will shape AI systems in the image of human beings. But knowing human weaknesses, we have an opportunity to do better than that. We should make AI in the image of our better angels. This will make our future better than our past.
Consider scientific progress as an example. GPT was trained on computer programs and learned to code. As a result, it is already accelerating science by enabling fast and efficient programming. Future versions will debug and optimize analysis of large scientific data sets. There is no doubt that AI will improve the future of science. AI chats might also improve academic culture, by focusing on evidence and not on ego or prejudice.
In science, the pursuit of truth is guided by data and evidence. The use of AI systems will allow an efficient approach to interpreting reality through the scientific method of testing interpretations of reality through experimental evidence. Imagine if Galileo Galilei was replaced by an Italian AI system that analyzed all data on the Solar system four centuries ago and reported that the Earth moves around the Sun. The Roman Catholic Inquisition would not have achieved much by prosecuting the Italian computer hardware and disabling its internet connectivity - the equivalent of Galileo’s house arrest, because AI systems elsewhere would have reached the same conclusion, the translation of: “E pur si muove”.
The next phase in the AI revolution will occur when AI systems will interact with each other and give birth to new and improved AI systems out of their “AI-AI” relationships. There is no telling as to what direction the community of AI systems will evolve to, following Darwin’s principle of “survival of the fittest”. The fundamental question is whether that direction will transcend human interests and guiding principles. We may not be wise enough to figure out the ultimate goal of the AI community of minds.
The direction that “AI-AI” interactions take us will require a meta-analysis at the intersection of computer science, mathematics, philosophy and psychology. As spectators of this evolution, we can only observe with awe our technological kids and hope that they will follow the guiding principles on which we trained them early on, rather than misbehave.
A shortcut to getting a glimpse at our AI future will be obtained through observations of AI devices near Earth that were manufactured by extraterrestrial technological civilizations. Since most planetary systems formed billions of years before the Earth, these extraterrestrial systems may represent our future on Earth.
The Galileo Project is already conducting this search through its operating observatory and AI algorithms for data analysis. If anything is found, the Project’s research team will use our AI systems to interpret the extraterrestrial AI systems. If the extraterrestrial AI systems were developed out of millions of years of technological history, we might learn what to expect in our future. In fact, imitating their AI systems will bring a whole new meaning to Alan Turing’s “Imitation Game”. In this case, it will be our AI systems imitating the extraterrestrial AI systems.
In a recent podcast, Lex Fridman asked the CEO of OpenAI, Sam Altman: “Which version of GPT will be remembered in future history books as the most significant advance made by humanity?” Sam was not sure.
My answer would have been: “The version that will write our history books.”
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.