Showing posts with label Friendly artificial intelligence. Show all posts
Showing posts with label Friendly artificial intelligence. Show all posts

Sunday, November 26, 2017

Why Superintelligent AI Could Be The Last Human Invention

Via ValueWalk.com,


When we create something more intelligent than we could ever be, what happens after that? We have to teach it.



MAX TEGMARK: Hollywood movies make people worry about the wrong things in terms of super intelligence. What we should really worry about is not malice but competence, where we have machines that are smarter than us whose goals just aren’t aligned with ours. For example, I don’t hate ants, I don’t go out of my way to stomp an ant if I see one on the sidewalk, but if I’m in charge of this hydroelectric dam construction and just as I’m going to flood this valley with water I see an ant hill there, tough luck for the ants. Their goals weren’t aligned with mine and because I’m smarter it’s going to be my goals, not the ant’s goals, that get fulfilled. We never want to put humanity in the role of those ants.


On the other hand it doesn’t have to be bad if you solve the goal alignment problem. Little babies tend to be in a household surrounded by human level intelligence as they’re smarter than the babies, namely their parents. And that works out fine because the goals of the parents are wonderfully aligned with the goals of the child’s so it’s all good. And this is one vision that a lot of AI researchers have, the friendly AI vision that we will succeed in not just making machines that are smarter than us, but also machines that then learn, adopt and retain our goals as they get ever smarter.


It might sound easy to get machines to learn, adopt and retain our goals, but these are all very tough problems. First of all, if you take a self-driving taxi and tell it in the future to take you to the airport as fast as possible and then you get there covered in vomit and chased by helicopters and you say, “No, no, no! That’s not what I wanted!” and it replies, “That is exactly what you asked for,” then you’ve appreciated how hard it is to get a machine to understand your goals, your actual goals.


A human cabdriver would have realized that you also had other goals that were unstated because she was also a human and has all this shared reference frame, but a machine doesn’t have that unless we explicitly teach it that. And then once the machine understands our goals there’s a separate problem of getting them to adopt the goals. Anyone who has had kids knows how big the difference is between making the kids understand what you want and actually adopt your goals to do what you want.


And finally, even if you can get your kids to adopt your goals that doesn’t mean they’re going to retain them for life. My kids are a lot less excited about Lego now than they were when they were little, and we don’t want machines as they get ever-smarter to gradually change their goals away from being excited about protecting us and thinking of this thing about taking care of humanity as this little childhood thing (like Legos) that they get bored with eventually.


If we can solve all three of these challenges, getting machines to understand our goals, adopt them and retain them then we can create an awesome future. Because everything I love about civilization is a product of intelligence. Then if we can use machines to amplify our intelligence then we have this potential to solve all the problems that are stumping us today and create a better future than we even dare to dream of.


If machines ever surpass us and can outsmart us at all tasks that’s going to be a really big deal because intelligence is power.



The reason that we humans have more power on this planet than tigers is not because we have larger muscles or sharper claws, it’s because we’re smarter than the tigers. And in the exact same way if machines are smarter than us it becomes perfectly plausible for them to control us and become the rulers of this planet and beyond.


When I. J. Good made this famous analysis of how you could get an intelligence explosion, or intelligence just kept creating greater and greater intelligence leaving us far behind, he also mentioned that this super intelligence would be the last invention that man need ever make. And what he meant by that, of course, was that so far the most intelligent being on this planet that’s been doing all the inventing - it’s been us.


But once we make machines that are better than us at inventing, all future technology that we ever need can be created by those machines if we can make sure that they do things for us that we want and help us create an awesome future where humanity can flourish like never before.









Saturday, November 11, 2017

When A.I. Rules...

Elon Musk unveiled his apocalytpic vision of the world a few weeks ago...


“Until people see robots going down the street killing people, they don’t know how to react because it seems so ethereal,” he said.


 


“AI is a rare case where I think we need to be proactive in regulation instead of reactive. Because I think by the time we are reactive in AI regulation, it’s too late.”


 


“Normally the way regulations are set up is a while bunch of bad things happen, there’s a public outcry, and after many years a regulatory agency is set up to regulate that industry,” he continued.


 


“It takes forever. That, in the past, has been bad but not something which represented a fundamental risk to the existence of civilization. AI is a fundamental risk to the existence of human civilization.”



And since then numerous futurists have prognosticated on whether is mankind"s salvation or eventual downfall. Facebook"s Mark Zuckerberg embraces it while Stephen Hawking considers this the most dangerous moment in history as AI and automation are set to decimate jobs and change the social contract.



However, as Mike Wehner via BGR.com,  writes, when AI rules, one rogue programmer could end the human race...


The idea of small groups of humans having control over some of the most powerful weapons ever to be built is scary, but it’s the reality we live in. In the not-so-distant future, that incredible power and responsibility could be handed over to AI and robotic systems, which are already in active development. In a pair of open letters to the prime ministers of bother Australia and Canada, hundreds of AI researchers and scientists are pleading for that not to happen.


The fear, they say, is that removing the human element from life and death decisions could usher in a destructive age that ultimately spells the end of mankind. The AI weapons systems are, as the researchers put it, “weapons of mass destruction” which must be banned outright before they can do any serious damage.


“Delegating life-or-death decisions to machines crosses a fundamental moral line – no matter which side builds or uses them,” the letter explains.


 


“Playing Russian roulette with the lives of others can never be justified merely on the basis of efficacy. This is not only a fundamental issue of human rights. The decision whether to ban or engage autonomous weapons goes to the core of our humanity.”



In a setting where computers have the ultimate say in whether or not to engage in hostile acts — even under the guise of defending their own territories or protecting the populations they are programmed to protect — conflicts could escalate much faster than humans have ever seen. Weeks, months, or even years of posturing and diplomacy could turn into mere minutes or even seconds, with missiles flying before humans can even begin to intervene. And then, of course, there’s the issue of the AI being manipulated in unforeseen ways.


“These will be weapons of mass destruction,” the scientists say.


 


“One programmer will be able to control a whole army. Every other weapon of mass destruction has been banned: chemical weapons, biological weapons, even nuclear weapons. We must add autonomous weapons to the list of weapons that are morally unacceptable to use.”



It’s a frightening thought, but it hasn’t stopped military contractors from exploring the possibility of AI-controlled weapons and defense systems. This could be yet another way mankind engineers its own destruction.


*  *  *








Saturday, October 14, 2017

Robocops Are Coming To London

All of Elon Musk’s worst fears are slowly becoming reality.


The Telegraph reports that robocop-type policing techniques are set to replace beat cops within a decade as the use of artificial intelligence becomes more widespread in investigating crimes.


Thames Valley Police said AI computers – which mimic humans by making decisions themselves – could be used to answer 999 calls, detect crimes and identify offenders.



But in one chilling detail that will hopefully be resolved within the next decade or so before robocops become a routine presence in London’s streets, the TVP warned of “bias” in the AI software and a concern that AI computers “might be unable to reason with a human”.


We don’t find further clues about the subject of these biases until much later in the story, when the leader of a civil liberties group – which have been predictably alarmed by plans to automate police forces – pointed to a study showing the machine’s algorithms have shown racial biases, according to the Telegraph.





Civil liberties groups were alarmed by the plans. David Green, the director of the Civitas thinktank, warned that the AI computers could unfairly target ethnic minority groups.



He said: “Robocop policing has now arrived in England. This Orwellian reliance on automated decisions has been found to undermine the most basic precepts of the justice system when it has been tried in America.



“An experiment in Fort Lauderdale, for example, found that the algorithm reflected human prejudices, including racial bias.”



However, the news comes as ministers prepare to publish the first ever review into how AI will change Britain over the coming decades. AI is already used by Scotland Yard to recognise faces at London’s Notting Hill Carnival. Durham Constabulary is also planning to use AI for deciding whether to keep suspects in custody.



In a submission to a Parliamentary inquiry into the Implications of Artificial Intelligence, Thames Valley said there are “even at the lowest level AI could perform many of the process driven tasks that take place in the police”.


AI could be used to assist “investigations by ‘joining the dots’ in police databases, the risk assessment of offenders, forensic analysis of devices, transcribing and analysis of CCTV and surveillance, security checks and the automation of many administrative tasks”, it said.


To be sure, Thames Valley police noted that even once robocops are ready for widespread use, they will still require "a high level of human oversight and clear justification".


Their report showed that “recent tests of AI in policing indicate there is a risk of bias perpetuation in AI outputs, therefore engagement with Privacy and Civil Rights groups will be necessary to persuade the public that everything possible is being done to mitigate this whilst doing our best to keep them safe.





“Of utmost importance is that any AI process that involves an ethical issue, must have a high level of human oversight and clear justification.



The automation of processes also introduces a risk of being unable to reason with a human when events occur outside expected parameters.”



Dubai earlier this year became the first major city to deploy robocops, saying that it expects them to comprise 25% of the United Arab Emirates’ police force by 2030.


Musk has repeatedly warned that humans are underestimating the dangers of AI, claiming that these technologies warrant strict government oversight so that robots don’t transform into a threat to humanity.


…just in case you forgot



Sunday, August 13, 2017

Elon Musk Doubles Down On AI Scare: "Artificial Intelligence Vastly More Risk Than North Korea"

With the world"s attention focused on the Korean Peninsula and the growing threat of global thermonuclear war, Tesla CEO Elon Musk has bigger things to worry about. In a series of "alarming" tweets on Friday, Musk warned the world should be more worried about the dangers of artificial intelligence than North Korea.


Having unveiled his apocalytpic vision of the world a few weeks ago...





“Until people see robots going down the street killing people, they don’t know how to react because it seems so ethereal,” he said.



“AI is a rare case where I think we need to be proactive in regulation instead of reactive. Because I think by the time we are reactive in AI regulation, it’s too late.”



“Normally the way regulations are set up is a while bunch of bad things happen, there’s a public outcry, and after many years a regulatory agency is set up to regulate that industry,” he continued.



“It takes forever. That, in the past, has been bad but not something which represented a fundamental risk to the existence of civilization. AI is a fundamental risk to the existence of human civilization.”



Musk was quickly admonished by another Silicon Valley billionaire as Mark Zuckerberg suggested Musk is exaggerating, noting:





“I have pretty strong opinions on this. I am optimistic. I think you can build things and the world gets better. But with AI especially, I am really optimistic.



“And I think people who are naysayers and try to drum up these doomsday scenarios — I just, I don"t understand it. It"s really negative and in some ways I actually think it is pretty irresponsible.”



And while the feud grows...



Musk took to Twitter to turn the fearmongery amplifier to "11"...





“If you"re not concerned about AI safety, you should be. Vastly more risk than North Korea,” he tweeted.



“Nobody likes being regulated, but everything (cars, planes, food, drugs, etc.) that"s a danger to the public is regulated. AI should be too,”






His stark warning came at a time when the US and North Korea remain on heightened alert amid spiraling tensions on the Korean Peninsula. Earlier this week, both sides degenerated to open threats, demonstrating readiness to use coercive force if provoked to do so.



But Musk appeared to be more frightened by artificial intelligence, a rising phenomenon he is willing to put under control.


We look forward to Mr. Zuckerberg"s response.