Should You Worry About the Global Pursuit for AI Domination?

• Updated on: 2017-09-06
• Read original article here

Should You Worry About the Global Pursuit for AI Domination?
Because one thing is for certain: the race to achieve it is certainly on.
Russian state news organization RT reports that Vladimir Putin believes "artificial intelligence is the future, not only for Russia, but for all humankind." In fact, he went a step further, adding that it comes "with colossal opportunities, but also threats that are difficult to predict. Whoever becomes the leader in this sphere will become the ruler of the world."
The state news agency also reports that Putin wouldn't abuse that kind of power. "If we become leaders in this area, we will share this know-how with entire world," he's reported to have said, "the same way we share our nuclear technologies today." You can probably take that part with a pinch of salt.
Meanwhile, China is rather more forward about its own intentions for achieving AI dominance. Earlier this year, the nation's government announced plans to surpass Western nations in the field and grow a machine learning industry worth $150 billion. (Incidentally, analysis by Goldman Sachs suggests that it's on track .)
All of which, predictably, has AI doomsayer Elon Musk a little flustered. Taking to Twitter, the Tesla and SpaceX CEO voiced his concern: "competition for AI superiority at [a] national level [is the] most likely cause of WW3 [in my opinion]." He later added that conflict "may be initiated not by the country leaders, but one of the AIs, if it decides that a pre-emptive strike is most probable path to victory."
All of which sounds rather ominous. So, the big question here really is: should you worry about AI's existential threat to humankind in the first place? Here are two views from MIT Technology Review: maybe , maybe not .
Source:
Should You Worry About the Global Pursuit for AI Domination?
Because one thing is for certain: the race to achieve it is certainly on.
Russian state news organization RT reports that Vladimir Putin believes "artificial intelligence is the future, not only for Russia, but for all humankind." In fact, he went a… Read more
Because one thing is for certain: the race to achieve it is certainly on.
Russian state news organization RT reports that Vladimir Putin believes "artificial intelligence is the future, not only for Russia, but for all humankind." In fact, he went a step further, adding that it comes "with colossal opportunities, but also threats that are difficult to predict. Whoever becomes the leader in this sphere will become the ruler of the world."
The state news agency also reports that Putin wouldn't abuse that kind of power. "If we become leaders in this area, we will share this know-how with entire world," he's reported to have said, "the same way we share our nuclear technologies today." You can probably take that part with a pinch of salt.
Meanwhile, China is rather more forward about its own intentions for achieving AI dominance. Earlier this year, the nation's government announced plans to surpass Western nations in the field and grow a machine learning industry worth $150 billion. (Incidentally, analysis by Goldman Sachs suggests that it's on track .)
All of which, predictably, has AI doomsayer Elon Musk a little flustered. Taking to Twitter, the Tesla and SpaceX CEO voiced his concern: "competition for AI superiority at [a] national level [is the] most likely cause of WW3 [in my opinion]." He later added that conflict "may be initiated not by the country leaders, but one of the AIs, if it decides that a pre-emptive strike is most probable path to victory."
All of which sounds rather ominous. So, the big question here really is: should you worry about AI's existential threat to humankind in the first place? Here are two views from MIT Technology Review: maybe , maybe not .
Source:
\n","length":337}">
Human Embryo Editing Study Shows We Still Have a Lot to Learn About CRISPR
The first human embryos edited in the U.S. appear to have had a faulty gene repaired—but now a debate is raging as to what actually happened.
In late July, MIT Technology Review broke the story about the work , in which researchers edited about 150 early-stage… Read more
The first human embryos edited in the U.S. appear to have had a faulty gene repaired—but now a debate is raging as to what actually happened.
In late July, MIT Technology Review broke the story about the work , in which researchers edited about 150 early-stage embryos using the CRISPR gene-editing technique. In the subsequent paper published in Nature, the team revealed that it was able to successfully eliminate a genetic mutation that causes a deadly heart condition. Importantly, the results suggested the edits occurred with a far higher level of precision than anyone else had managed before in embryos. One of the study's authors, Shoukhrat Mitalipov, talked of clinical trials being near at hand .
But questions have emerged this week about how, exactly, the faulty gene was removed. The authors of the original study claim the gene was repaired as CRISPR cut out only the faulty DNA, which was on the paternal side, then used normal maternal DNA as a template to correct the mutation—a previously unknown phenomenon. ( Watch our CRISPR explainer )
On Monday, a different group of researchers  called this account into question  in a paper posted on bioRxiv. They pointed out that Mitalipov's team only showed that the faulty gene is absent from embryos after editing, not that the gene had been repaired. What's more, paternal and maternal DNA are still distinct in the embryos the team was using. So how could the two have interacted?
Instead, the cross-examiners suggest—and many have subsequently agreed —that it's possible that CRISPR could have been making much larger deletions of the embryos' DNA. If that's true, then the faulty gene would fail to show up when Mitalipov's team went looking for it, but the embryos could have a great deal of genetic damage besides. Without ruling out this possibility—or else figuring out another way to avoid so-called "off-target" effects —it would be irresponsible to suggest that CRISPR-edited embryos be implanted and allowed to grow into children.
Mitalipov has responded to the criticism by encouraging scientists to perform their own experiments to confirm his team's findings—an admirably scientific way to both stand by one's work and acknowledge that the case is far from closed. In the meantime, those clinical trials may be on hold for a while longer yet.
Source:
\n","length":381}">
Houston Is Swarming with Drones
The skies above the flooded city are packed with unmanned vehicles performing a range of tasks. AT&T is using drones to inspect its cellular towers for damage, while insurance companies like Allstate and Farmers are rolling out their own fleets to follow… Read more
The skies above the flooded city are packed with unmanned vehicles performing a range of tasks. AT&T is using drones to inspect its cellular towers for damage, while insurance companies like Allstate and Farmers are rolling out their own fleets to follow up on claims—as of Wednesday, Farmers already had 14,000 claims relating to Hurricane Harvey, according to the San Antonio Express .
This is likely the first disaster in which many companies are deploying drones in large numbers, but it fits with a growing trend in the insurance business . Instead of sending human adjusters to inspect every claim that customers make, more and more firms are sending drones. Improvements in image recognition mean that software is playing a bigger role in assessing how much damage a car sustained in an accident, say, or how much the roof of a house needs to be repaired after a storm.
Rescue operations are benefitting, too. According to Axios , the company DroneDeploy is sending out vehicles to produce detailed 3-D maps that can help navigate the watery chaos. The company claims it can speed up rescue operations by providing imagery that allows rescuers to see around buildings and beneath tree cover. 
Source:
\n","length":311}">
Would a Universal Basic Income Be Good for the Economy?
A study from the Roosevelt Institute has concluded exactly that . It suggests that a government handout to every American of $12,000 a year, no strings attached, would boost the U.S. economy to the tune of a cumulative 12.5 to 13 percent over eight years.… Read more
A study from the Roosevelt Institute has concluded exactly that . It suggests that a government handout to every American of $12,000 a year, no strings attached, would boost the U.S. economy to the tune of a cumulative 12.5 to 13 percent over eight years. The report also says it would create more jobs.
Okay, time for the massive grain(s) of salt: the report's view is highly controversial, for a couple of reasons. First, the economic model it uses assumes that the U.S. economy is sitting in a state of artificially low demand. One reason for this, some economists think, is that income inequality has led a few rich people to hoard a disproportionate amount of wealth, instead of spending it, which acts to have a chilling effect on economic activity. (This is a view most notably espoused by Thomas Piketty—see " Technology and Inequality .") Assuming that's the case, a huge cash infusion from the government might be worth it because, the model suggests, it would lead to economic growth worth around $2.5 trillion, as well as the creation of millions of jobs. As Vox points out in its analysis of the report , though, this is an opinion that many economists disagree with.
A second assumption is potentially even more misleading. The Roosevelt Institute assumes that people given a basic income would not be inclined to work less. They cite several small-scale studies that support this. But as we have written before in an in-depth look at the impact a universal basic income would have in the U.S. , the evidence is murky at best. That makes such a scheme incredibly risky—both from a labor standpoint and, if the return in economic activity is no guarantee, for its overall pricetag.
Source:
\n","length":324}">
Meet Hyperloop’s Chinese Competitor
One Chinese aerospace firm wants to blow away anything that Elon Musk has proposed with something it calls a "flying train." From Quartz :
The country’s state-run space contractor, China Aerospace Science and Industry Corporation (CASIC), announced yesterday… Read more
One Chinese aerospace firm wants to blow away anything that Elon Musk has proposed with something it calls a "flying train." From Quartz :
The country’s state-run space contractor, China Aerospace Science and Industry Corporation (CASIC), announced yesterday (Aug. 30) that it has started research on a “high-speed flying train” that it says will be able to reach top speeds of 4,000 km per hour (2,485 miles per hour). That is 10 times faster than the world’s fastest bullet train (which is also in China), four times faster than commercial flights, and over three times the speed of sound (1,225 km/h).
Much like a hyperloop, CASIC's flying train would use magnetic levitation and travel through a tube that's had the air sucked out of it to reduce drag. Claims that the train will exceed Mach 3 should be met with some skepticism—no terrestrial vehicle has ever come close to that.
But in general, high-speed transit through evacuated tunnels is starting to look like it could actually work. Earlier this month, Hyperloop One (which Musk isn't involved with)  managed to fire a pod through a tunnel  at nearly 200 miles per hour. And late last night Musk posted on Instagram that SpaceX's own pod topped 220 mph and might try for 310 mph next month.
That's still far short of the goal Musk set when he first described the idea of a hyperloop of traveling around the speed of sound, but both Hyperloop One and SpaceX have now built prototypes and are rapidly progressing through tests. That's several steps ahead of CASIC's effort—for the moment, at least.
Source:
\n","length":316}">
AI Lets Astrophysicists Analyze Images 10 Million Times Faster
If you’re ever casually analyzing the cosmological phenomenon that is gravitational lensing in a hurry, you’re best off using neural networks. That’s certainly what researchers from SLAC National Accelerator Laboratory and Stanford University found:… Read more
If you’re ever casually analyzing the cosmological phenomenon that is gravitational lensing in a hurry, you’re best off using neural networks. That’s certainly what researchers from SLAC National Accelerator Laboratory and Stanford University found: their analysis of the distortions in spacetime using AI are 10 million times faster than the methods they used to use.
Gravitational lensing is the effect that’s observed when a massive object in space, like a cluster of galaxies, bends light that’s emitted from, say, a more distant galaxy. When observed by telescopes, it causes distortions in images—and analysis of those distortions can help astronomers work out the mass of the object that caused the effect. And, perhaps, even shed a little light on the distribution of dark matter in the universe.
The problem: comparing recorded images to simulations of gravitational lenses used to take weeks of human effort. Now, writing in Nature , the team explains that it’s built neural networks that are trained to recognize different lenses, by studying half a million computer simulations of their appearance. Turned on real images, the AI can work out what kind of lens—and therefore the type of mass—that affected the observed light as well as human analysis, but almost instantly.
“Analyses that typically take weeks to months to complete, that require the input of experts and that are computationally demanding, can be done by neural nets within a fraction of a second, in a fully automated way,” said Stanford’s Laurence Perreault Levasseur in a statement. “And, in principle, on a cell phone’s computer chip.”
The technology underlying this sort of AI image recognition has become increasingly common in many applications over recent years , from social networks spotting faces to search engines recognizing objects in photographs. But scientists demand utmost rigor, and while neural networks have been applied to astrophysics problems before, according to a statement by Stanford’s Roger Blandford, they have they have done so  “with mixed outcomes.”
Now, says Blandford, there’s “considerable optimism that this will become the approach of choice for many more data processing and analysis problems in astrophysics.” 
Source: