Is artificial intelligence a threat to human civilization? Let’s take a look at AI safety

[[125155]]

Kurzweil has written five books on artificial intelligence (AI), one of which is the New York Times bestseller How to Create a Mind.

Two great thinkers saw the dangers of AI. Here's how to make it safe.

Renowned physicist Stephen Hawking[1] recently warned that if artificial intelligence surpasses human intelligence, it will pose a threat to the existence of human civilization. Elon Musk, a pioneer in digital currency, private space flight, and electronic cars, has expressed similar concerns.

If AI becomes an existential threat, it won’t be the first. As a child, I sat under a table during a 1950 civil-raid drill. Since then, we’ve encountered comparable fears, such as bioterrorists creating a new virus that humans have no resistance to. Fire can warm us or burn our villages; technology is always a double-edged sword.

In typical dystopian futuristic movies, there are always one or two protagonists or groups fighting for control of artificial intelligence. AI is being integrated into the world today. AI is not one or two hands; it is a billion or two billion hands. An African child with a smartphone has access to more knowledge than the president of the United States 20 years ago. As AI gets smarter, its uses are still growing. In fact, everyone's intelligence has been amplified by AI in the last 10 years.

We still have conflict among people, each amplified by AI, as has been the case. But we can take comfort in the fact that violence is declining profoundly, exponentially, as Steven Pinker’s 2011 book, The Better Angels of Our Nature: Why Violence Is Declining, suggests. According to Steven Pinker, while statistics vary by location, the death rate from war is hundreds of times lower than it was six centuries ago. Murders have fallen tenfold since then. People are astonished. The impression that violence is on the rise stems from another trend: the exponentially greater abundance of negative information about the world—another development enabled by AI.

There are strategies we can deploy to keep emerging technologies like AI safe. Consider biotechnology, which is perhaps decades ahead of AI. The 1975 Asilomar Conference on recombinant DNA assessed its potential dangers and mapped out strategies to keep the field safe. Since then, the best practices have been refined by industry, and they work remarkably well: there have been no major problems, accidental or intentional, in the past 39 years. We are now seeing tremendous progress in drug therapies reaching clinical practice, with far fewer problems than expected.

The ethics of AI can be traced back to Isaac Asimov’s 1942 short story “Runaround”[2]’s three laws of robotics[3], and eight years later, Alan Turing’s 1950 paper “Computing Machinery and Intelligence” on AI. The common view among AI practitioners today is that we are still decades away from achieving human-level AI. I am more optimistic and put the time at 2029, but there is still time to plan ethical standards.

There are already efforts by universities and companies to promote AI safety strategies and guidelines, and some are already on track. Similar to the guidelines of the Asilomar conference, one idea is to clearly define the mission of each AI program and establish cryptographic security protections to block unauthorized users.

Ultimately, the most important way we can ensure AI safety is through commitment to human governance and social regulation. We have already reached the age of human-machine civilization. The best way to avoid negative conflicts in the future is to continue to develop our social ideals, which greatly reduce violence.

Today, AI is advancing disease diagnosis, finding cures, developing renewable clean energy, helping clean the environment, delivering high-quality education to people around the world, helping people with disabilities (including Stephen Hawking's voice), and contributing in countless other ways. In the coming decades, we have the opportunity to make a giant leap in addressing humanity's great challenges. In controlling this danger, we have a moral responsibility to deliver on that promise. This will not be the first time we succeed in this regard.

• Note 1: Stephen Hawking; OCH, CBE, FRS, FRSA (English: Stephen Hawking, January 8, 1942 -), a famous British physicist, is the director of research at the Center for Theoretical Cosmology at the University of Cambridge. Hawking has made many important contributions: he and Roger Penrose jointly proposed the Penrose-Hawking theorem in the framework of general relativity, and he theoretically predicted that black holes will emit radiation, now known as Hawking radiation. http://zh.wikipedia.org/wiki/%E5%8F%B2%E8%92%82%E8%8A%AC%C2%B7%E9%9C%8D%E9%87%91
• Note 2: "Runaround" is a short science fiction story written by American writer Isaac Asimov, which contains characters Powell and Donovan who appear repeatedly in many of his subsequent novels. The story was written in October 1941 and first published in the March 1942 issue of Super Science Stories magazine. http://zh.wikipedia.org/wiki/%E7%8E%AF%E8%88%9E_(%E7%9F%AD%E7%AF%87%E5%B0%8F%E8%AF%B4)
• Note 3: The Three Laws of Robotics are the behavioral rules set by science fiction writer Isaac Asimov for robots in his robot-related works and other robot-related novels. It is another famous fictional theory of Asimov besides "Psychohistory". http://zh.wikipedia.org/wiki/%E6%9C%BA%E5%99%A8%E4%BA%BA%E4%B8%89%E5%AE%9A%E5%BE%8B

Original English text: http://time.com/3641921/dont-fear-artificial-intelligence/

Translation from: http://www.labazhou.net/2014/12/dont-fear-artificial-intelligence/