The Benefits of a Successful Singularity

\r\nGetting artificial intelligence right could lead to a new world of abundance.\r\nPart five in a GOOD miniseries on the singularity by Michael...

Getting artificial intelligence right could lead to a new world of abundance.

Part five in a GOOD miniseries on the singularity by Michael Anissimov and Roko Mijic. New posts every Monday from November 16 to January 23.

"...the first ultraintelligent machine is the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control."

I.J. Good, "Speculations Concerning the First Ultraintelligent Machine" (1963)

In a post three weeks ago, Roko Mijic explained why singularity researchers are so interested in the prospect of smarter-than-human intelligence. The intelligence of Homo sapiens which we all share underlies all the achievements of each individual human and the collective human family. Human intelligence underlies our artistic creations, philosophical treatises, humanitarian efforts, scientific reasoning, and much more. Chimps cannot participate in any of these areas of thought and action, because their intelligence is not sufficient.

If human researchers end up successfully creating general intelligence in a machine, the event would constitute our first encounter with a true "alien intelligence" that is our equal. This article will assume that the programmers of the first artificial intelligences have successfully solved the extremely difficult problems of both describing intelligence in terms of algorithms and ensuring that the resulting minds are explicitly human-friendly. If they succeeded, what would be the benefits?

One of the greatest limitations of human intelligence is its speed. Our neurons fire about 200 times per second under ideal conditions. In contrast, silicon transistors can send and receive electronic signals more than 2 billion times per second. Our brains are massively parallel, with about a hundred billion neurons, but each neuron operates so slowly that a $10,000 desktop supercomputer can execute 933 billion operations per second, which is within a factor of 100,000 of one of the most common estimates of the processing power of the human brain, 10^17 operations per second. Several supercomputers have already passed the 10^15 operations per second mark.

Most artificial intelligence researchers agree that the primary challenge is not hardware, but software. If we understood the processes of intelligence in enough detail to implement them on computers, it seems likely that ample hardware would be available to create many copies. These copies could also be run in an accelerated fashion on distributed systems. This "instant intelligence, just add computing power" principle is a recipe for explosive economic growth.

Artificial intelligences with human interests in mind could use their machine minds to solve human challenges in fields like medicine, physics, chemistry, engineering, politics, diplomacy, biology, sociology, and economics. Being native to the world of computers, they could run complex simulations in mere moments that would take human researchers years to build. Complex, detailed, mathematically accurate simulations could be the default thought mode for artificial intelligences-their "thoughts" could be far superior to our best simulations.

Recent findings by researchers at EPFL in Lausanne, Switzerland determined that perceptual learning-like learning how to detect anomalies on medical images that are extremely hard to detect for untrained people-can proceed just as effectively though mental imagery (imagination) as through exposure to real data. This research suggests that there is a threshold quantity of information that needs to be soaked up in a given domain, after which significant gains might be made simply through advanced internal simulation, analysis, and "practice." This bodes well for the prospect of AIs making concrete contributions to the economy even if they initially lack physical bodies.

Having direct access to all the freely available data and information on the internet, combined with complex internal simulations, and eventually accompanied by direct experiments through robotic proxies, the first roughly human-equivalent artificial intelligences could contribute huge value to humanity within a very short time. Humanity is presently faced with a glut of information: We have more data than we know what to do with. Artificial intelligence will be the key to studying that data and generating novel theories completely beyond human inferential capabilities. Combined with "everything machines" of the type I described two weeks ago, this could lead to tremendous amounts of material wealth for humanity. The entire planet could be raised to the standard of living of the West or better, all on clean power plants built and maintained by automatic systems.

The question is not whether we will gain immense material wealth and knowledge. That seems likely if we create AI. The question is whether our wisdom will keep pace with these material gains, and whether we will be able to avoid conflict in a new world of abundance. A civilization with immense power and knowledge still has to decide which goals to direct that power and knowledge towards-and these are not easy questions. Maybe some combination of human and machine intelligence will bring us closer to the answers.

Michael Anissimov is a futurist and evangelist for friendly artificial intelligence. He writes a Technorati Top 100 Science blog, Accelerating Future. Michael currently serves as Media Director for the Singularity Institute for Artificial Intelligence (SIAI) and is a co-organizer of the annual Singularity Summit.

AFP News Agency / Twitter

A study out of Belgium found that smart people are much less likely to be bigoted. The same study also found that people who are bigoted are more likely to overestimate their own intelligence.

A horrifying story out of Germany is a perfect example of this truth on full display: an anti-Semite was so dumb the was unable to open a door at the temple he tried to attack.

On Wednesday, October 9, congregants gathered at a synagogue in Humboldtstrasse, Germany for a Yom Kippur service, and an anti-Semite armed with explosives and carrying a rifle attempted to barge in through the door.

Keep Reading Show less
via Andi-Graf / Pixabay

The old saying goes something like, "Possessions don't make you happy." A more dire version is, "What you own, ends up owning you."

Are these old adages true or just the empty words of ancient party-poopers challenging you not to buy an iPhone 11? According to a new study of 968 young adults by the University of Arizona, being materialistic only brings us misery.

The study examined how engaging in pro-environmental behaviors affects the well-being of millenials. The study found two ways in which they modify their behaviors to help the environment: they either reduce what they consume or purchase green items.

Keep Reading Show less

One of the biggest obstacles to getting assault weapons banned in the United States is the amount of money they generate.

There were around 10 million guns manufactured in the U.S. in 2016 of which around 2 million were semiautomatic, assault-style weapons. According to the National Shooting Sports Foundation, the firearms industry's trade association, the U.S. industry's total economic impact in 2016 alone was $51 billion.

In 2016, the NRA gave over $50 million to buy support from lawmakers. When one considers the tens of millions of dollars spent on commerce and corruption, it's no wonder gun control advocates have an uphill battle.

That, of course, assumes that money can control just about anyone in the equation. However, there are a few brave souls who actually value human life over profit.

Keep Reading Show less
via Reddit and NASA / Wikimedia Commons

Trees give us a unique glimpse into our past. An examination of tree rings can show us what the climate was like in a given year. Was it a wet winter? Were there hurricanes in the summer? Did a forest fire ravage the area?

An ancient tree in New Zealand is the first to provide evidence of the near reversal of the Earth's magnetic field over 41,000 years ago.

Over the past 83 million years there have been 183 magnetic pole reversals, a process that takes about 7,000 years to complete.

Keep Reading Show less
The Planet
via Pixabay

The final episode of "The Sopranos" made a lot of people angry because it ends with mob boss Tony Soprano and his family eating at an ice cream parlor while "Don't Stop Believin'" by Journey plays in the background … and then, suddenly, the screen turns black.

Some thought the ending was a dirty trick, while others saw it as a stroke of brilliance. A popular theory is that Tony gets shot, but doesn't know it because, as his brother-in-law Bobby Baccala said, "You probably don't even hear it when it happens, right?"

So the show gives us all an idea of what it's like to die. We're here and then we're not.

Keep Reading Show less