George Rebane
‘Scientists Worry Machines May Outsmart Man’ is the title of an uneven yet revealing piece in the 25jul09 NYT by John Markoff. The article was apparently motivated by a quiet (‘secret’ sounds too sinister) conference at Asilomar on Monterey Bay sponsored by the Association for the Advancement of Artificial Intelligence last February. It was a quiet conference in comparison to the concurrent announcements at the time that preceded the opening of the Singularity University this summer. RR’s Singularity Signposts section has kept readers up to date on this and other events that herald the Singularity as the epochal new age of Man.
The February conference was lead by Microsoft’s Dr. Eric Horvitz who also happens to be the current president of AAAI. According to the NYT, the conference was motivated by a growing cohort of well-read people from many areas who were now “impressed and alarmed by advances in artificial intelligence”. It focused the discussion of “a group of computer scientists … debating whether there should be limits on research that might lead to loss of human control over computer-based systems that carry a growing share of society’s workload, from waging war to chatting with customers on the phone.”
The attendees seemed to agree that “there was legitimate concern that technological progress would transform the work force by destroying a widening range of jobs, as well as force humans to learn to live with machines that increasingly copy human behaviors. … The researchers also discussed possible threats to human jobs, like self-driving cars, software-based personal assistants and service robots in the home.”
The central idea of the Singularity is the ongoing “intelligence explosion” as proclaimed by Dr. Ray Kurzweil and others. The inevitable result of this explosion is a world in which ever smarter machines will design ever smarter machines – and this explosion is already well on its way. The result, for the instant we call the Singularity, will be the arrival of peer intelligence (‘what if your computer were as smart as your spouse?’), and then we enter the unknown post-Singularity era. All the rules will change.
What everyone (again, only the well-read need apply) is now sensing was summarized by Dr. Horvitz, “Something new has taken place in the past five to eight years, … , technologists are replacing religion, and their ideas are resonating in some ways with the same idea of the Rapture.” More on these themes is available here at AAAI.
But as RR readers know, this epochal transformation is currently apprehended only by a miniscule fraction of humanity (more on that below). One of the functions of the Singularity University is preparation. This summer the institution began courses to train a “cadre” of people who will attempt “to shape the advances and help society cope with the ramifications.” Lots of luck.
Later this year the AAAI will issue a report of these proceedings. The report will consider and assess various futures including “ ‘the loss of human control of computer-based intelligences.’ It will also grapple, Dr. Horvitz said, with socioeconomic, legal and ethical issues, as well as probable changes in human-computer relationships.” How would it be to interact with a machine that is both sentient (conscious) and sapient (smart, wise)? And what roles in society would such machines assume?
As we await this report, I invite you to consider the following short list of observations, notions, and apprehensions.
1. Larger and larger segments of human labor will be performed better and more cheaply by machines.
2. Governments will start controlling the possession of general purpose computers (the kind we now have on our desks and laptops). Only ‘dumb client’ machines will be legal in the hands of the general public.
3. Intelligent machines will spread and operate in the nether world of rogue technorati and criminals.
4. Governments will start passing set-aside labor laws that will seek to preserve certain jobs/tasks to be performed only by humans.
5. The pre-Singularity years will witness luddite riots of unimaginable proportions that start in the developed world and spread to the less developed countries.
6. Keeping sentient and sapient machines operating in vacuo (not connected to the realworld) will be attempted by all governments. The attempts will be futile as these machines (let’s call them artificial beings) leak out into the world’s communication and control grids (e.g. the Internet).
7. Obtaining Machine-augmented-Man (MAM) functionality will become the highest goal of most/all humans. Soon, humans without MAM implants, parts, and/or genetic modifications will be de facto second class citizens wherever they reside.
8. Religions that cannot embrace the imminent Singularity will initially join the luddites, and ultimately disappear. Religions that can, will survive to become the main cultural and cognitive links to the pre-Singularity era.
9. All reproduction will be controlled to the extent that the controlling classes will only produce ‘designer progeny’.
10. Eugenics in various forms will eventually become the national policy of all advanced countries as they seek to ‘manage their natural populations’.
Finally, the attendees at the conference offered sanguine predictions that we will not achieve a global super-intelligence, at least any time soon. Given the history of such wide-ranging pronouncements from leading scientists over the ages, that prediction scares hell out of me.
===
Personal Disclosure – I was privileged to have contributed to the advancement of machine learning, early causality, and reasoning (e.g. Rebane, Pearl 1998) when doing part of my doctoral research more than twenty years ago. Staying current in the literature is now a losing proposition. However, the paths that we all then pioneered have grown into broad and well-travelled avenues with many productive side streets and intriguing alley ways. Every day it becomes more and more clear that machine sentience and sapience will be nurtured and not designed. The tragedy may come if such a machine, unintended by humans, spontaneously evolves into being.
As a life-long inventor with a list of patents under my belt, I was early on educated in the arena of luddite economic behavior. In the 1960s, working day and night, my brother-in-law and I invented and built a utensil sorter for institutional and large restaurant kitchens. It was a marvelous device that had only one moving part (US Patent #3389790). The country’s two largest kitchen systems manufacturers both wanted to manufacture the machines under license. We selected one and looked forward to an unending flow of royalty checks.
The first sorters were bought and installed into the in-flight meal preparation kitchens of one of the world’s largest airlines located at San Francisco International. Soon, with the first royalty check in hand, we took our (long suffering and saintly) wives out to a celebratory dinner – you can imagine the grand evening that we had after all the years of work. Well, it was to be the first and last. About two weeks later we received word from the manufacturer that the airline was taking the sorters out of service.
It turned out that the machines were being sabotaged by union instructed employees who purposely jammed all the sorters with thousands of knives, forks, and spoons at a rate much faster than they could sort. The airline was not going to fight the union in the attempt to reduce the kitchen's sorting staff by more than two thirds. The luddites won that round hands down. Chastened by reality – for no good deed goes unpunished – and sadder but wiser, we all went on with our lives.
George,
A very interesting view of our future. I started working on Intelligent Vehicle Highway Systems (IVHS) in the early 1990s. TRW provided the electronic steering systems in six Dodge vehicles to the PATH Program of the University of California Berkeley, funded by the State of California Business, Transportation, and Housing Agency Department of Transportation (CALTRANS); and the
United States Department of Transportation, Federal Highway Administration. The goal was fully automate vehicles on a specially prepared highways. The concept was platooning, using radar to connect the vehicles into vehicle trains on special highway lanes.
PATH demonstrated the vehicles on special section of I-15 with an HOV lane. In the morning and evening the track carried inbound and out bound traffic to San Diego, the rest of the day it was an automate vehicle test track. Being a grant driven program, it had a lot of starts and stops due to funding limitations. I did not give automate platoons much hope, because the program failed to recognize how people view technology and liability issues. When crap happens, who pays the bill? Caltrans? Insurance? Vehicle Owner? State? Federal Government? No one stepped up!
As we become more and more dependent on robots, and crap happens, who pays the bill? We are a very litigant society, everything that happens to us is someone else's fault. Who will defend the robots in court? Who pays the bill, when the court determines damages. Will we get a fair trial if the judge is a robot? Way too many questions. Way too scary, one day we may wake up and find the robots are replicating themselves and no human has control of the power switch. I am with you, the academics are clueless when it comes to assessing reality.
Posted by: Russ Steele | 26 July 2009 at 07:33 PM