If we want AI to explain itself, here’s how it should tell us

\n","length":359}">
If we want AI to explain itself, here’s how it should tell us
Explainable AI systems aim to make decisions that are easily understood by humans—a laudable goal, but what makes a good explanation?
Testing the best: There’s only one way to figure that out: ask some users. So that’s what researchers from Harvard and Google Brain did, in a series of studies. Test subjects looked at different combinations of inputs, outputs, and explanations around a machine learning algorithm that was designed to learn the dietary habits or medical conditions of an alien (Yes, seriously—alien life was chosen to avoid the test subject’s own biases creeping in). Users then scored the different combinations.
Keep it short: Longer explanations were found to be more difficult to parse than shorter ones—though breaking up the same amount of text into many short lines was somehow better than making people read a few longer lines. As you can tell, the tests examined some pretty basic elements of how to deliver information—but at least it’s a start.
Image credit:
\n","length":358}">
The UK says Russia was behind the huge NotPetya ransomware attack
In a rare example of directly attributing blame, the British government says Russia orchestrated the massive cyber attack in 2017.
Backstory: Last summer a new breed of ransomware, dubbed NotPetya and based on a Windows flaw leaked from the NSA, held… Read more
In a rare example of directly attributing blame, the British government says Russia orchestrated the massive cyber attack in 2017.
Backstory: Last summer a new breed of ransomware, dubbed NotPetya and based on a Windows flaw leaked from the NSA, held computers around the world hostage . It hit Ukraine particularly hard, but was felt globally—from India’s largest container port to US hospitals.
Blaming Russia:  According to the Guardian , the UK's foreign office minister for cyber security, Lord Ahmad, says that "the Russian government, specifically the Russian military, was responsible for the destructive NotPetya cyber-attack."  
Bracing for more: Meanwhile, UK defense secretary Gavin Williamson says that "we have entered a new era of warfare, witnessing a destructive and deadly mix of conventional military might and malicious cyber-attacks .. We must be primed and ready to tackle these stark and intensifying threats.”
Source:
\n","length":337}">
Google Chrome now blocks ads—but it may be biased
While the search firm has built an ad blocker into its browser, some reports suggest that it’s a self-serving exercise.
What's blocked: As of today, Chrome blocks ads that don't meet standards laid out by the Coalition for Better Ads . Think pop-ups and… Read more
While the search firm has built an ad blocker into its browser, some reports suggest that it’s a self-serving exercise.
What's blocked: As of today, Chrome blocks ads that don't meet standards laid out by the Coalition for Better Ads . Think pop-ups and auto-play audio on your desktop, or flashing animations and countdowns on your mobile.
But: The Wall Street Journal reports that some industry experts are raising concerns about the blocks. “Google overly influenced the process that selected which ad types to block,” writes the newspaper, and may have skewed things in a way that could help it turn even more of a profit from advertising.
Why it matters: The ad blocking is being billed by Google as a useful service, but, if the reports are true, its motivations may be off. Regulators, especially anti-competition powers in the EU that are already out for Google’s blood, will be paying close attention.
Source:
\n\n\t\t
\n","length":1792}">
“We’re in a diversity crisis”: cofounder of Black in AI on what’s poisoning algorithms in our lives
Artificial intelligence is an increasingly seamless part of our everyday lives, present in everything from web searches to social media to home assistants like Alexa. But what do we do if this massively important technology is unintentionally, but fundamentally,...
Read the full story →
Artificial intelligence is an increasingly seamless part of our everyday lives, present in everything from web searches to social media to home assistants like Alexa. But what do we do if this massively important technology is unintentionally, but fundamentally, biased? And what do we do if this massively important field includes almost no black researchers? Timnit Gebru is tackling these questions as part of Microsoft’s Fairness, Accountability, Transparency, and Ethics in AI group, which she joined last summer. She also cofounded the Black in AI event at the Neural Information Processing Systems (NIPS) conference in 2017 and was on the steering committee for the first Fairness and Transparency conference in February. She spoke with MIT Technology Review about how bias gets into AI systems and how diversity can counteract it.
\n","length":359}">
If we want AI to explain itself, here’s how it should tell us
Explainable AI systems aim to make decisions that are easily understood by humans—a laudable goal, but what makes a good explanation?
Testing the best: There’s only one way to figure that out: ask some users. So that’s what researchers from Harvard and… Read more
Explainable AI systems aim to make decisions that are easily understood by humans—a laudable goal, but what makes a good explanation?
Testing the best: There’s only one way to figure that out: ask some users. So that’s what researchers from Harvard and Google Brain did, in a series of studies. Test subjects looked at different combinations of inputs, outputs, and explanations around a machine learning algorithm that was designed to learn the dietary habits or medical conditions of an alien (Yes, seriously—alien life was chosen to avoid the test subject’s own biases creeping in). Users then scored the different combinations.
Keep it short: Longer explanations were found to be more difficult to parse than shorter ones—though breaking up the same amount of text into many short lines was somehow better than making people read a few longer lines. As you can tell, the tests examined some pretty basic elements of how to deliver information—but at least it’s a start.
Image credit:
\n","length":331}">
China’s AI startups scored more funding than America’s last year
Of $15.2 billion invested in AI startups globally in 2017, 48 percent went to China and just 38 percent to America. So says a new report from CB Insights about the state of AI.
So long, America: It’s the first time China’s AI startups surpassed those… Read more
Of $15.2 billion invested in AI startups globally in 2017, 48 percent went to China and just 38 percent to America. So says a new report from CB Insights about the state of AI.
So long, America: It’s the first time China’s AI startups surpassed those in the US in terms of funding. While America still has more AI startups than China, they’re starting to lose out in striking equity deals: the US accounted for 77 percent of them in 2013, but that fell to 50 percent last year.
Fierce competition: AI startup investment rose 141 percent in 2017 compared with 2016—but with 1,100 new startups appearing last year, AI appearing in business models everywhere, and Big Tech’s enterprise AI offerings gaining traction, it’s harder than ever to snag funds.
Why it matters:  China is all-in on AI , but the US government is doing little to support its own industry . The success of Chinese startups is another sign that America could soon be left behind.
Image credit:
\n","length":319}">
Delivery drones could reduce emissions—if they’re used properly
Regulating how drones are used to haul cargo could ensure that they help fight global warming.
Size is important: A Lawrence Livermore study published in Nature Communications shows that a small quadcopter drone carrying a one-pound load from a local… Read more
Regulating how drones are used to haul cargo could ensure that they help fight global warming.
Size is important: A Lawrence Livermore study published in Nature Communications shows that a small quadcopter drone carrying a one-pound load from a local warehouse emits less carbon dioxide than a medium-duty truck would to deliver the same package.
So is power: Things are less clear for a large octocopter drone hauling 16 pounds. If the electricity used to power the drone is generated from fossil fuels, the drone creates 50 percent more emissions than the truck. If it comes from renewables, the drone creates 9 percent less.
Why it matters: Joshua Stolaroff, who led the study, tells IEEE Spectrum that this is a rare chance to “consider the impacts of the technology before it gets deployed.” That means delivery drone use could be limited to ensure that it results in fewer emissions, rather than more.
Source:
\n","length":315}">
Here are the big tech threats worrying US intelligence officials
Heads of America’s federal intelligence agencies met before the Senate Tuesday to discuss the biggest threats facing the world. Here are the tech troubles that the FBI, CIA, National Security Agency, and Office of the Director of National Intelligence… Read more
Heads of America’s federal intelligence agencies met before the Senate Tuesday to discuss the biggest threats facing the world. Here are the tech troubles that the FBI, CIA, National Security Agency, and Office of the Director of National Intelligence are worrying about right now:
Election meddling. “Persistent and disruptive cyber operations will continue against the United States and our European allies using elections as opportunities to undermine democracy,”  explained Dan Coats, director of national intelligence .
IP gobbling. “[China is] exploiting the very open R&D environment we have, which we all revere,” said FBI director Christopher Wray . “They’re taking advantage of it.” The level of infiltration by Chinese students and researchers in the US, he said, showed a “level of naïveté” in academia.  
More hacks. Russian attacks on power stations, Chinese digital espionage, North Korean cyber-heists to raise funds—Coats warned that there’s plenty more to come .
Image credit: