Stephen Hawking recently made headlines by noting, "The development of full artificial intelligence could spell the end of the human race." Others, conversely, have trumpeted a new age of "superintelligence" in which smart devices will exponentially extend human capacities. No longer just a matter of science-fiction fantasy (2001, Blade Runner, The Terminator, Her, etc.), it is time to seriously consider the reality of intelligent technology, many forms of which are already being integrated into our daily lives. In that spirit, John Brockman, publisher of Edge.org ("the world's smartest website" —The Guardian), asked the world's most influential scientists, philosophers, and artists one of today's most consequential questions: What do you think about machines that think?
Steven Pinker considers the internal metal life of robots * Frank Tipler explains how artificial intelligence (AI) will save humanity and colonize space * Martin Rees explores why humans are merely an evolutionary stage on the path to a machine-dominated world * Nicholas Carr examines the challenges of maintaining control over machines * Daniel C. Dennett identifies the true danger of the coming technological "singularity" * Nobel Prize winner Frank Wilczek asserts that all intelligence is machine intelligence * musician Brian Eno suggests that human society remains our most powerful supercomputer * George Dyson argues that genuine creative thinking will always be analog, not digital * Alison Gopnik asks whether machines will ever be as smart as a three-year-old * Richard Thaler thinks human stupidity will always impede artificial intelligence * Wired founder Kevin Kelly calls AIs an "alien intelligence" * plus contributions from Nobel laureate John C. Mather, Matt Ridley, Freeman Dyson, Douglas Rushkoff, Helen Fisher, Sam Harris, George Church, Hans Ulrich Obrist, Esther Dyson, Nick Bostrom, and others.
Philip Tetlock: If you turn to session six, slide 117-118, you’re going to see a little piece on the seductive power of scenarios. Imagine you’ve got one of these between subjects designs in which half of the people read the top slide, half of the people read the bottom slide, then they make a judgment about the plausibility or probability of this outcome.
Philip Tetlock: What I want to do today is three things: I want to clear up some confusion from last time about counterfactual reasoning and how it’s intertwined with superforecasting; I want to link up some of the things that are in the final sets of slides on sacred values and taboo cognition and superforecasting; Then, I want to give you some examples of superforecasting in action and talk about condensing it all into four big problems and one killer app solution.
Philip Tetlock: I was thinking over lunch about how to summarize or capture the essence of what went on this morning. There were certain key ideas I wanted to get across and for the most part they’ve gotten across, but some may have been miscommunicated a bit based on some feedback I’ve gotten.
You can look at the glass as either one-third full or two-thirds empty. Forecasting tournaments are in their infancy as a scientific method and as a tool for improving policy debate. We’ve gone through the first generation of tournaments, and we’ve made tangible progress. We’ve learned how to keep score; we have shown that it’s possible to measure the accuracy of probabilistic judgments, of messy real world events; we have shown that it’s possible to improve accuracy through a combination of selecting the right people, training them, teaming them, and using the right types of aggregation algorithms. Those are all achievements, but we’re still far short of what I’m going to call the “Bob Axelrod ideal” of a mechanism that can guide policy with evidence based precision. I don’t know if that’s a fair characterization of where you would want us to be, but we’re at least two-thirds short of that. We have made some progress, and we’re on the right path.
Philip Tetlock: I thought it might be a good idea before we go into the quality of questions that are input into tournaments to say a few more words about the nature of the Good Judgment Project and what it meant to win the forecasting tournament, how it won the forecasting tournament, and what inferences you might or might not want to draw from victory in forecasting tournaments in general.
If you were to turn to slide twenty-eight in the book, it raises the question, how much can the Good Judgment Project improve foresight?
For the psychology professor Philip Tetlock, the hunt for Osama Bin Laden is a classic example of the insufficiency of secret-service agencies. When Barack Obama gave the green light for that operation four years ago, he knew he was making one of the most difficult decisions in his life—one that would not only mean life or death for those involved, but also sway the course of history and help determine his legacy. The prognoses offered by the secret-service agencies were inconclusive: some put the likelihood for success at 40%, others at 80%. In the movie based on this operation, Zero Dark Thirty, the CIA agent Maya insists she is 100% certain of success. In reality, Obama determined the chances stood at fifty-fifty and gave the green light against the advice of his secretary of defense.
In Tetlock's view, such imprecisions present an unacceptable risk. Forecasts alleging complete certainty are, of course, unscientific. But Tetlock argues that a historic decision must not be based on imprecise reports. While Obama may have enjoyed luck on a historic scale, with his special task force finding Bin Laden and killing him, Tetlock insists that the work of secret-service agencies must change—fundamentally.
Since the eighties Tetlock has worked on precisely this endeavor. For four years now he has pursued research at the University of Pennsylvania at the behest of the Intelligence Advanced Research Projects Activity (IARPA), which the NSA and the CIA, together with fourteen other American secret-service agencies, established in 2006, in order to develop new methods for secret-service work in the post-9/11 era. Among IARPA’s divisions are the Office for Anticipating Surprise, the Office of Smart Collection, and the Office of Incisive Analysis.
Psychologists' "forecasting tournaments" capture the interest of the NSA and the CIA
This past weekend Tetlock met with twenty scientists and engineers on a vineyard north of San Francisco. Two European journalists were invited; otherwise, the meeting was closed to the public. Tetlock wanted to discuss the results of his Good Judgment Project, which he has worked on for 24 years. The scientists discuss the project under ideal circumstances: sheltered from the summer heat in the cool living room of a stately Victorian house. With palms in the garden, a front porch and wainscoting, the house exudes colonial splendor. The air is redolent with the rose beds in front of the windows and the precious woods of the furniture. The host is John Brockman of Edge Foundation, Inc. (http://edge.org), which is the best network for such debates in the country. That explains the presence of such intellectual heavyweights as the Nobel Laureate in Economics Daniel Kahneman, the political scientist and National Medal of Science winner Robert Axelrod, the political scientist Margaret Levi, and Google Vice President Salar Kamangar. It isn’t easy to hold one’s own in such a group. Kahneman in particular, the cleverest of them all, is skeptical.
Tetlock begins by recounting the history of the Good Judgment Project. In 1984 he began holding "forecasting tournaments" in which selected candidates are asked questions about the course of events. In the wake of a natural disaster, what policies will be changed in the United States? When will North Korea test nuclear weapons? Candidates examine the questions in teams. They are not necessarily experts, but attentive, shrewd citizens. One of the best forecasters so far is Bill Flack, a former official of the U.S. Department of Agriculture from Nebraska.
This can't be the end of human evolution. We have to go someplace else.
It's quite remarkable. It's moved people off of personal computers. Microsoft's business, while it's a huge monopoly, has stopped growing. There was this platform change. I'm fascinated to see what the next platform is going to be. It's totally up in the air, and I think that some form of augmented reality is possible and real. Is it going to be a science-fiction utopia or a science-fiction nightmare? It's going to be a little bit of both.
JOHN MARKOFF is a Pulitzer Prize-winning journalist who covers science and technology for The New York Times. His most recent book is the forthcoming Machines of Loving Grace: The Quest for Common Ground Between Humans and Robots. John Markoff's Edge Bio Page
THE NEXT WAVE
I'm in an interesting place in my career, and it's an interesting time in Silicon Valley. I grew up in Silicon Valley, but it's something I've been reporting about since 1977, which is this Moore's Law acceleration. Over the last five years, another layer has been added to the Moore's Law discussion, with Kurzweil and people like him arguing that we're on the brink of self-aware machines. Just recently, Gates and Musk and Hawking have all been saying that this is an existential threat to humankind. I simply don't see it. If you begin to pick it apart, their argument and the fundamental argument of Silicon Valley, it's all about this exponential acceleration that comes out of the semiconductor industry. I suddenly discovered it was over.
Now, it may not be over forever, but it's clearly paused. All the things that have been driving everything that I do, the kinds of technology that have emerged out of here that have changed the world, have ridden on the fact that the cost of computing doesn't just fall, it falls at an accelerating rate. And guess what? In the last two years, the price of each transistor has stopped falling. That's a profound moment.
On the weekend of July 30th, Edge convened one of its "Master Classes." In the past, these classes have featured short courses taught by people such as psychologist and Nobel Laureate Daniel Kahneman ("A Short Course in Thinking About Thinking"); behavioral economists Richard Thaler and Sendhil Mullainathan, again with Kahneman ("A Short Course in Behavioral Economics"); and genomic researchers George Church and J. Craig Venter ("A Short Course on Synthetic Genomics").
This year, the psychologist and social scientist Philip E. Tetlock presented the findings based on his work on forecasting as part of the Good Judgment Project. In 1984, Tetlock began holding "forecasting tournaments" in which selected candidates were asked questions about the course of events: In the wake of a natural disaster, what policies will be changed in the United States? When will North Korea test nuclear weapons? Candidates examine the questions in teams. They are not necessarily experts, but attentive, shrewd citizens.
Steven Pinker, who has written about Tetlock's work on Superforecasting, noted that "Tetlock is one of the very, very best minds in the social sciences today. He has come up with one brilliant idea after another, and superforecasting is no exception. Everyone agrees that the way to know if an idea is right is to see whether it accurately predicts the future. But which ideas, which methods, which people have an actual, provable track record of non-obvious predictions vindicated by the course of events? The answers will surprise you, and have radical implications for politics, policy, journalism, education, and even epistemology—how we can best gain knowledge about the world we live in."
Among Tetlock's "students" at the Edge weekend were many intellectual heavyweights including political scientist and National Medal of Science winner Robert Axelrod; psychologist, Nobel Laureate, and recipient of the 2013 Presidential Medal of Freedom Daniel Kahneman; the political scientist and Director of Stanford’s CASBS Margaret Levi; Google Senior Vice President Salar Kamangar; psychologist and National Medal of Science winner Anne Treisman; Roboticist Rodney Brooks, former head of MIT's Computer Science Lab; W. Daniel Hillis, pioneer in massively parallel computation; medical inventor Dean Kamen; and Peter Lee, Corporate Vice President, Microsoft Research, overseeing MSR NExT.
Over the weekend in Napa, Tetlock held five classes, which are being presented by Edge in their entirety (8.5 hours of video and audio) along with accompanying transcripts (61,000 words). Commenting on the event, one of the participants wrote:
"The interesting thing is that this is not about a latest trend that might scale in one or two years, but about real change that might take a decade or two. Also, these masterclasses are not only much more profound than any of the conferences popularizing contemporary intellectualism. The possibility to spend that much time with the clairvoyants in a setting like this also gives you a sense of community so much greater than any of the advertised."
PHILIP E. TETLOCK, Political and Social Scientist, is the Annenberg University Professor at the University of Pennsylvania, with appointments in Wharton, psychology and political science. He is co-leader of the Good Judgment Project, a multi-year forecasting study, the author of Expert Political Judgment and (with Aaron Belkin) Counterfactual Thought Experiments in World Politics, and co-author (with Dan Gardner) of Superforecasting: The Art & Science of Prediction (forthcoming, US, Crown, September 29th; UK, Random House, September 24th). Further reading on Edge: "How To Win At Forecasting: A Conversation with Philip Tetlock" (December 6, 2012). Philip Tetlock's Edge Bio Page.