You hear that? It’s the “Jaws” theme music playing, so beware. Yes, Henry Kissinger is back. Actually, at age 98, he has never left the world scene. He’s not my cup of tea, to say the least. That said, we at the Council on Foreign Relations (CFR) just spent an interesting hour with him on the subject of Artificial Intelligence (AI). Not bad for any nonagenarian.
Dr. Kissinger was joined by former Google CEO and co-founder, Dr. Eric Schmidt, who’s the real AI expert in that pairing. Together they have authored with Dr. Daniel Huttenlocher the recently released book, The Age of AI: And Our Human Future. The virtual session was moderated by PBS’ Judy Woodruff.
First, though, a humorous CFR aside. Kissinger was addressing the organization decades ago when another egomaniac entered the room late and in animated fashion for all to see as Kissinger was speaking. The late Richard Holbrooke bellowed, “Oh don’t pay any attention to me, Henry” whereupon the acerbic, quick-witted Kissinger immediately retorted, “I never have." Touché!
Kissinger told us that AI is a tool for both good and bad. He worries deeply about its implications in national security and statecraft; thus his involvement in the book project. Of course, as with most innovations, the answer is that AI will serve both beneficial and destructive purposes. For example, Schmidt spoke of the drug Halicin. Developed by MIT researchers using in silico deep-learning AI, Halicin is an antibiotic of unprecedented value that could not have been created by humans alone. On the other hand, Schmidt asked how parents will feel when their children’s best friends are AI stuffed animals that are smarter than any human and teach kids things we would rather they not know at such tender ages. Kissinger said that “we will be altering the human experience” but that society has far more questions than answers on the AI subject.
Teddy bears aside, there is a race among nations – starting with China and the U.S. – to achieve AI competitive advantage. The nation with the best AI capabilities at the right time could, for example, detect perceived military or cyber threats otherwise undetectable by human-led systems today and activate offensive actions, even if the threats were not real or worth first-mover advantage. Hello, WarGames (1983). “How about a nice game of chess?” Kissinger rightly drew parallels to the protocols he and others established in the 1950s to stabilize the race to develop nuclear weapons. Schmidt added that, “There are no treaties on AI” and that the nuclear protocols were created only after the tragedies of Hiroshima and Nagasaki. “I’d like to see us do something before a massive tragedy occurs,” he said. Unfortunately, that’s not how the human species works.
Kissinger said, “AI will change the conception of human reality” and that studying its consequences is crucial.” Schmidt flatly said about all of this, “We’re playing with fire.” He spoke of AI systems in the near future that will be instructed to choose which longstanding problems to solve in physics, for example, and keep learning all along the way. “They will choose which problems to work on, solve them, and you won’t know how they did it. “It’s a new epoch,” he said, “because we will no longer be the top thinkers.” Then what happens? “Do we try to kill it? Do we invent a religion around it?” Schmidt provocatively but properly asked. Okay, you can name the most applicable Star Trek episode at your convenience.
This is why Schmidt and Kissinger argue for a holistic approach to this challenge with philosophers, ethicists, educators, sociologists, economists, and many other disciplines participating in the AI conversation. Technologies rarely serve society well when left only in the hands of technologists. “We can’t have computer scientists solely in charge of this,” Schmidt warned. You sure you don’t want to play that nice game of chess?
