ConceptNet, an artificial intelligence program developed by a team led by Catherine Havasi at the MIT Media Lab, performed as well as an average four-year-old on the information, vocabulary, and word reasoning portions of standard intelligence test. The program uses a crowdsourced semantic network—a database of statements of basic facts—to answer questions. Miles Brundage explained in Slate that the program did well on “precisely the parts of the test that one would expect computers to excel at,” but that it did poorly compared to a four-year-old on the comprehension portion of the test. Brundage wrote that
things we humans think of as hard (playing chess, winning Jeopardy!) have been conquered by researchers, but things we think of as easy (recognizing a chess board and moving pieces around it without knocking other ones over, walking around the set of Jeopardy!) remain unsolved.
For that reason, the program’s success may not say much about how close we are to developing computer programs that can compete with human intelligence.
A World Meteorological Organization report found that the decade from 2001-2010 was the warmest on record for both land and ocean surfaces in both hemispheres. The US Energy Information Administration projected in its latest energy outlook report that world energy consumption will increase 56% between 2010 and 2040. The report estimated that energy-related carbon dioxide emissions would increase 46% if there were no change to current regulations. A study in Ecology Letters concluded that adapting to climate change over the next 100 years would require rates of niche evolution that are unprecedented among vertebrates. The paper found that under current projections the climate will change 10,000 times faster than most vertebrates can adapt to environmental changes and that species that are unable to move to cooler areas will have a hard time surviving.
In a comment in Nature, Gail Whiteman, Chris Hope, and Peter Wadhams used a standard model of the costs of climate change to calculate that methane—a potent greenhouse gas—released as the arctic permafrost melts could do $60 trillion worth of damage to the global economy. Although the release of arctic methane will likely be felt around the world, they estimated that the world’s less-developed economies will bear 80% of the costs. While any estimate of the costs of climate change is necessarily speculative, if their calculation is correct the costs would be almost as much as the world economy produces in a whole year. They didn’t account for the effects of ocean acidification or changing atmospheric circulation, so the total economic impact of climate change in the Arctic could be even higher.
The US Central Intelligence Agency is funding a $630,000 study through the National Academy of Sciences on the risks and effectiveness of geoengineering the climate. The CIA has studied the possibility of controlling the weather since the Vietnam War, when it seeded clouds to rain on the Ho Chi Minh Trail. Because some geoengineering proposals are relatively inexpensive, individual states or even wealthy individuals might be able to alter the environment unilaterally. Last year, Russ George, an American entrepreneur, dumped 120 metric tons of particulate iron into the Pacific Ocean in order to cause a bloom of plankton that would absorb carbon dioxide from the atmosphere. Right now there is very little domestic or international regulation of geoengineering experiments.
The US threatened to cancel President Obama’s trip to Moscow in the fall if Russia grants asylum to NSA-leaker Edward Snowden, although Obama still plans to attend the G20 summit in St. Petersburg. The threat probably also reflects tensions over Russian support of Syrian leader Bashar al-Assad and concerns that a trip to Moscow might not accomplish much. Russian President Vladimir Putin suggested granting Snowden asylum wouldn’t be worth it, saying “it’s like shearing a pig: there’s lots of squealing and little fleece.”
Russia’s deputy defense minister Anatoly Antonov proposed updating conventional weapons treaties that date back to the Cold War. New weapons, he said, had erased much of the difference between nuclear and conventional weapons. New treaties should take into account the effectiveness of drones, space-based weapons, and modern missile-defense systems. A new web application called NukeMap 3D uses Google Earth to simulate the effects of a nuclear strike over different locations. The application allows you to estimate casualties for detonations of different sizes at different locations. And the US Department of Defense denied a Freedom of Information Act request from the Federation of American Scientists for documents containing information about the current size of the US nuclear arsenal. The Defense Department denied the request in spite of having revealed in 2009 the number of warheads the US had at the time. The Defense Department said then that “increasing the transparency of global nuclear stockpiles is important to non-proliferation efforts.”
A study in PLoS One suggested that pesticides may be to blame for some of the deaths of bees used in agriculture. The study found 35 different pesticides in pollen collected from bees in five different states across the country. Some of the pesticides were at concentrations that were high enough to kill large numbers of bees. Researchers particularly noted the presence of fungicides that increase the bees’ susceptibility to parasites. Neonicotinoid insecticides, which have been banned in Europe because of evidence they may play a role in bee deaths, were not found in most of the pollen samples.
Researchers developed vaccines that protect monkeys against both rabies and the Ebola virus. The success of the vaccines indicates that it might be possible to develop a similar vaccine for humans. Ebola kills between 80-90% of humans who contract the disease. A paper in Nature modeled the spread of smallpox after a deliberate release of the virus by terrorists. The paper concluded that given the amount of international travel there is the virus would probably spread to several different countries before health systems became aware of the outbreak. And William Karesh and Catherine Machalaba wrote in The Huffington Post that pandemic diseases are potentially more than just a public health problem. A major pandemic could threaten the world’s food supplies, ecosystems, and economies. Right now we have identified only a small percentage of disease-causing agents and don’t have the infrastructure to detect and respond to dangerous outbreaks. Karesh and Machalaba argued that properly preparing for a pandemic emergence will mean investing billions of dollars a year in prevention, detection, and response programs.
The World Health Organization’s Emergency Committee said that the Middle East Respiratory Syndrome virus is “of serious and great concern” but does not at the moment meet the agency’s definition of a “Public Health Emergency of International Concern”. A Lancet study estimated that the basic reproductive number of the MERS virus, which represents the number of new cases each case is likely to generate, is just 0.69. Diseases with a basic reproductive number greater than 1 represent a pandemic risk because they are likely to spread, while diseases with a basic reproductive number less than one eventually stop spreading. The basic reproductive number of the pre-pandemic SARS virus, which is similar to the MERS virus, was around 0.80. The study found that there was a 5% chance the basic reproductive number of the MERS virus was greater than 1. There have been 90 confirmed MERS infections and 45 deaths.
Although the spread of H7N9 bird flu has slowed since June, a study in Science suggested that it may be transmissible in humans. The study showed that an H7N9 virus isolated from humans could readily be transmitted through the air among ferrets, which are often used to model disease transmission in humans. Other studies found that the virus has lower transmission rates than the seasonal flu, but that it reproduces quickly in human airway cells. A paper in mBio found new evidence that H7N9 can resist antiviral neuraminidase inhibitors like Tamiflu. Robert Webster, a corresponding author of the paper, said that we need to develop vaccines soon because “the antiviral option for controlling H7N9 isn’t too good.” There have been 134 confirmed H7N9 infections and 43 deaths.
Maryn McKenna argued in Nature that bacteria that are resistant to the otherwise broadly effective carbapenem class of antibiotics are becoming a serious health crisis. Carbapenem-resistant enterobacteria are responsible for serious blood, bladder, and lung infections and kill as many as half the patients who contract them. These bacteria have spread rapidly over the last 10 years and have been found in 18% of the hospitals that offer long-term critical care in the US. Sally Davies, the UK’s Chief Medical Officer, said in March that antibiotic resistance was as serious a threat as terrorism.
This news summary was put together in collaboration with and is cross-posted at Anthropocene. Thanks to Seth Baum, Kaitlin Butler, Linda Szabados, and Grant Wilson for help compiling the news.
For last month’s news summary, please see GCR News Summary June 2013.
You can help us compile future news posts by putting any GCR news you see in the comment thread of this blog post, or send it via email to Grant Wilson (grant [at] gcrinstitute.org).
Image credit: CSIRO.
