【人工智能威胁人类吗?谷歌又有新招了!】

【人工智能威胁人类吗?谷歌又有新招了!】谷歌DeepMind实验室和牛津大学准备研发一个“大红色按钮”来确保AI在人类的控制之中,只要AI表现出对人类的威胁,这个“大红色按钮”便能关掉系统,强制AI中断工作,而且能确保AI无法找到方法来阻止这个中断。那么这样是不是人工智能威胁论就不存在了呢?关键是AI真的无法阻止这个中断吗?

Google's AI team is developing 'big red button' to switch off systems if they pose a threat to humans

  • Researchers say AI agents are 'unlikely to behave optimally all the time' 
  • Machines could deliberately wipe out the human race or do so by misinterpreting an instruction, because our language is so subtle  

Google's secretive AI division is working on a 'big red button' that can halt its artificial intelligence software.

Researchers have previously warned that AI could threaten humanity, with doomsday scenarios of AIs taking over, with one expert involved in the new paper admitting Google's historic win over Go world champion proves AI can be 'unpredictable and immoral'.

Now the DeepMind team say they have the answer - an off switch.

Scroll down for video

2A16250F00000578-3143275-Dr_Armstrong_is_concerned_that_a_simple_instruction_to_an_artifi-a-27_1435587347308

The researchers, Laurent Orseau of Deep Mind and Dr Stuart Armstrong at the Future of Humanity Institute at Oxford University, explain in the paper that AI agents are 'unlikely to behave optimally all the time.'

'Now and then it may be necessary for a human operator to press the big red button to prevent the agent from continuing a harmful sequence of actions - harmful either for the agent or for the environment - and lead the agent into a safer situation.'

They say that 'safe interruptibility' is key.

'Safe interruptibility can be useful to take control of a robot that is misbehaving and may lead to irreversible consequences, or to take it out of a delicate situation, or even to temporarily use it to achieve a task it did not learn to perform or would not normally receive rewards for this.'

However, they admit that some systems could be unable to be halted.

'It is unclear if all algorithms can be easily made safely interruptible,' the authors admit.

Dr Armstrong has previously warned it's a race against time to develop safeguards around artificial intelligence research, before robots outwit us - or even accidentally cause our demise.

Speaking at a debate on artificial intelligence in London earlier this year, he warned that humans could be wiped out even if robots are instructed to protect people.

Dr Armstrong, based at the Future of Humanity Institute at Oxford University, predicted that robots will be increasingly integral to our everyday lives, doing menial tasks, but will eventually make humans redundant and take over,The Telegraph reported.

2A16251300000578-3143275-Dr_Armstrong_of_the_Future_of_Humanity_Institute_at_Oxford_Unive-a-26_1435587272597

He believes machines will work at speeds inconceivable to the human brain and will skip communicating with humans to take control of the economy and financial markets, transport, healthcare and more.

The robots will have what's known as artificial general intelligence (AGI), enabling them to do much more than carry out specific and limited tasks.

2A16251800000578-3143275-image-a-21_1435587048610

'Anything you can imagine the human race doing over the next 100 years there's the possibility AGI will do very, very fast,' he said.

Dr Armstrong is concerned that a simple instruction to an AGI to 'prevent human suffering' could be interpreted by a super computer as 'kill all humans' or that 'keep humans safe' could lead to machines locking people up.

'There is a risk of this kind of pernicious behaviour by an AI,' he said, adding that human language is subtle and can be easily misinterpreted.

'You can give AI controls, and it will be under the controls it was given. But these may not be the controls that were meant.'

He predicts that it will be difficult to tell whether a machine has deadly 'intentions' or not and could act as if it is a benefit to humanity right until the point it takes control of all functions.

Professor Stephen Hawking has previously told the BBC: 'The development of full artificial

2A16252000000578-3143275-Professor_Stephen_Hawking_pictured_has_previously_told_the_BBC_T-a-49_1435588160879

This echoes claims he made earlier in the year when he said success in creating AI 'would be the biggest event in human history, [but] unfortunately, it might also be the last.'

In November, Elon Musk, the entrepreneur behind Space-X and Tesla, warned that the risk of 'something seriously dangerous happening' as a result of machines with artificial intelligence, could be in as few as five years.

He has previously linked the development of autonomous, thinking machines, to 'summoning the

Speaking at the Massachusetts Institute of Technology (MIT) AeroAstro Centennial Symposium in October last year, Musk described artificial intelligence as our 'biggest existential threat'.

He said: 'I think we should be very careful about artificial intelligence. If I had to guess at what our biggest existential threat is, it's probably that. So we need to be very careful with artificial intelligence.

'With artificial intelligence we're summoning the demon. You know those stories where there's the guy with the pentagram, and the holy water, and … he's sure he can control the demon? Doesn't work out.'

While Dr Armstrong acknowledges that super intelligent computers could find cures for cancer and other illnesses, for example, but said that mankind is now in a race to create safe artificially intelligent machines, before it's too late.

24F165C500000578-0-image-a-22_1464993512572

One suggestion is to teach robots a moral code, but Dr Armstrong is pessimistic this will work because humans find it hard to separate right and wrong and are often not good role models when it comes to exemplary behaviour.

A group of scientists and entrepreneurs, including Elon Musk and Professor Hawking signed an open letter in January promising to ensure AI research benefits humanity.

The letter warns that without safeguards on intelligent machines, mankind could be heading for a dark future.

The document, drafted by the Future of Life Institute, said scientists should seek to head off risks that could wipe out mankind.

3226BD3F00000578-3624671-image-a-19_1464993373399

The authors say there is a 'broad consensus' that AI research is making good progress and would have a growing impact on society.

It highlights speech recognition, image analysis, driverless cars, translation and robot motion as having benefited from the research.

'The potential benefits are huge, since everything that civilisation has to offer is a product of human intelligence; we cannot predict what we might achieve when this intelligence is magnified by the tools AI may provide, but the eradication of disease and poverty are not unfathomable,' the authors writ

2A16252000000578-3143275-Professor_Stephen_Hawking_pictured_has_previously_told_the_BBC_T-a-49_1435588160879

But they issued a stark warning that research into the rewards of AI had to be matched with an equal effort to avoid the potential damage it could wreak.

For instance, in the short term, it claims AI may put millions of people out of work.

In the long term, it could have the potential to play out like a fictional dystopias in which intelligence greater than humans could begin acting against their programming.

'Our AI systems must do what we want them to do,' the letter says.

'Many economists and computer scientists agree that there is valuable research to be done on how to maximise the economic benefits of AI while mitigating adverse effects, which could include increased inequality and unemployment.'

链接:http://www.dailymail.co.uk/sciencetech/article-3624671/Google-s-AI-team-developing-big-red-button-switch-systems-pose-threat.html


Comments are closed.



无觅相关文章插件