With Air Pollution On the Rise, Researchers Sound Alarms


As the impeachment debate raged in Washington this week, the National Bureau of Economic Research released an important new working paper highlighting the subtler ways that negligent governance can influence thousands of lives.

In the paper, Carnegie Mellon researchers Karen Clay and Nicholas Muller estimate that a rise in fine particulate matter, or PM2.5, has led to 9,700 premature deaths in the United States in the last two years.

Elevated levels of PM2.5 contribute to heart disease, cancer, respiratory infections, and many other conditions. As Undark reported in a series last year, PM2.5 pollution is an entrenched global problem, killing millions of people each year. Levels of PM2.5 had been dropping in the United States for a decade, though, and the evidence of a sudden rise, compiled from hundreds of EPA monitoring stations across the country, is likely to provoke concerns.

Clay and Muller suggest that an increased use of natural gas, more vehicle travel, and large wildfires have all contributed to the uptick. They also find that decreased enforcement of air pollution regulations may have driven the rise in PM2.5 levels since 2016.

So is this finding a result of Trump Administration policies? The picture is complicated. The relevant enforcement actions began declining in 2013, under the Obama Administration. But they have continued under the Trump Administration, which has rolled back regulations intended to limit PM2.5 emissions, all part of a larger gutting of environmental regulations.

Nor is it likely that the Administration will take action to curb the public health threat. Indeed, the same day that the NEBR released its working paper, the Independent Particulate Matter Review Panel—a group of 20 experts who served as a EPA Clean Air Scientific Advisory Committee until October 2018, when, they write, they were “dismissed without notice by press release”—released its own evaluation of existing federal standards for PM2.5.

The 183 page-long report, addressed to EPA Administrator Andrew Wheeler, offers a clear warning. Current standards, the experts write, “are not protective of public health.”

Also in the news:

• Oil giant ExxonMobil went on trial Tuesday, accused of misleading investors about the financial risks it faces due to regulations designed to combat global warming. According to InsideClimate News, the historic case, brought by New York Attorney General Letitia James, is the “first major climate change lawsuit to reach trial in the United States.” Investigators allege that the Texas-based company produced different estimates for how much it would cost to comply with government regulations — one set it showed to investors and one that it released to the public. The internal version, lawyers for the attorney general argue, cited lower figures, suggesting that regulations would be less strict. This came after the company had already known — and tried to deny — for decades that fossil fuels were linked to global warming. While Exxon doesn’t dispute that it used two different cost estimates, it says it was transparent with investors. The trial is expected to continue for another two weeks, and its outcome could have a big impact on future climate lawsuits. (Inside Climate News)

• The Washington Post’s Drew Harwell took a deep dive into what technology experts are calling a “profoundly disturbing” new trend in job recruiting: Companies using artificial intelligence to make hiring decisions. The leading AI technology in this space, developed by the company HireVue, analyzes footage from webcam interviews to collect data on a job candidate’s facial expressions, vocal mannerisms, and speech patterns. From that data, it calculates an “employability score,” which can then be used to screen applicant pools. Harwell reports that more than a million job candidates have been analyzed by HireVue’s technology, and the company’s clients include major employers like Hilton, Unilever, and Goldman Sachs. But critics have been quick to point out the flaws and potential dangers of AI-driven hiring: AI algorithms are generally bad at reading facial expressions; they don’t explain their decisions; and HireVue’s technology is essentially designed to reward candidates who look and sound like a company’s existing high-performing workers, which could reinforce systemic biases. “It’s pseudoscience. It’s a license to discriminate,” AI researcher Meredith Whittaker told The Post. (The Washington Post)

• Melting glaciers have revealed five new islands in the Arctic, the largest of which is approximately the size of ten football fields, the Russian Defense Ministry announced this week. A Russian expedition mapped the islands in August and September, confirming a 2016 paper that first described the islands based on a review of satellite data collected by the Russian Navy. The rapid depletion of glacial ice led to the physical discovery of the new islands, according to expedition leader Aleksandr Moiseyev. Scientists are alarmed by the speed at which climate change is dramatically altering the landscape of the Arctic; this summer, 11 billion tons of ice disappeared in Greenland on a single day in August, following months of record high temperatures. And, in September, Antarctica’s third largest ice shelf, the Amery, lost a 345 billion-ton iceberg, nicknamed the “Molar Berg” for its tooth-like shape. Ice melt in the polar caps has the potential to accelerate: The melting ice creates a feedback loop in which even more ice melts at an even faster rate. Experts warn that if glaciers continue to melt at this rate, up to a fifth of the world’s population may be displaced by 2100 due to rising sea levels. (CNN)

• In response to the recent rash of school shootings, many schools are now turning to digital surveillance of their student’s communications, The Guardian reported this Tuesday. Technology companies such as Bark and Gaggle offer schools automated 24-hour surveillance of student’s school emails, chat messages, and shared documents. The tools send automated notifications to school officials when they detect certain “concerning phrases.” Guardian reporter Lois Beckett found multiples instances of school administrators intervening when they were notified that one of their students had written about self-harm. These programs do not shut down outside of school hours, and any students who communicate using a school email or messenger app at night will still have their messages under surveillance. While schools have informed students that their communications are being monitored, reportedly some students have been surprised by the speed with which school staff responded to their messages—sometimes within a matter of minutes. Privacy experts have pointed out that the constant monitoring of students may be harmful to their mental health and unfairly target minorities or those with disabilities, but the companies that offer these services counter that their products increase students’ safety. Bark states that it has helped prevent “16 credible school shootings”, whereas Gaggle asserts that it assisted in preventing up to 700 planned or attempted suicides. (The Guardian)

• In a study published this week in Nature, Google researchers announced that they had achieved quantum supremacy over classical computers using their 53 qubit quantum computer. Called Sycamore, the computer performed a simple operation in three minutes and 20 seconds that Google claims would take 10,000 years for a powerful, classical machine, IBM’s Summit supercomputer, to execute. But IBM challenged this analysis in a preprint published on their research blog two days before the Google paper was formally released (the study had been briefly leaked on a NASA site prior to official publication). The IBM team says that, with a different kind of classical computing technique, the Summit supercomputer could achieve the same result as Sycamore in only 2.5 days. There’s more at stake for IBM than the performance of its Summit machine; the company announced last month the imminent release of its own 53 qubit computer. (Nature

• Since Medicare first began covering 3D mammograms in 2015, the health insurance program has spent an additional $240 million on breast cancer screenings despite a lack of conclusive evidence that the new exams actually improve patient outcomes compared to traditional, two-dimensional mammograms, according to a Kaiser Health News (KHN) investigation published this week. The technology’s rapid adoption – the machines first hit the market in 2011, and an estimated 63 percent of U.S. mammogram facilities now have them – is due in part to the industry’s ferocious marketing efforts, which include paying influential doctors and researchers, enlisting celebrity cancer survivors as spokespeople, funding advocacy groups, and lobbying insurers to change their policies. Some patients told KHN they felt pressure from their healthcare providers to opt for a more expensive exam, while others were told it was a more accurate test, even though studies remain mixed on whether 3D exams find more tumors – and even if detecting more tumors actually improves patient safety. “It’s unethical to push a product before you know it helps people,” Otis Brawley, a cancer screening expert at John Hopkins University, told KHN. (Kaiser Health News)

• A new article in Nature detailed considerable advancements in CRISPR gene editing that could help correct 89 percent of the known disease-causing genetic variations in DNA. Currently, scientists use CRISPR to make cuts and insertions that reconfigure specific DNA sequences. The new techniques move closer to a “search-and-replace” function. Known as “prime editing,” the technique enables scientists to change any individual letter of a DNA sequence, or insert and delete entire sections of a DNA sequence. The tool has already been used to more efficiently and precisely repair gene mutations in human cells, including inserting sequences up to 44 letters and deleting sequences up to 80 lettersAccording to one of prime editing’s inventors, David Liu of the Broad Institute, this upgrade to CRISPR is capable of covering more than 75,000 DNA changes that are associated with genetic diseases. (STAT)

• And finally: Japan’s environment minister, Shinjiro Koizumi, apparently sent his fellow lawmakers into fits of righteous confusion recently when he argued publicly that a successful Japanese campaign to combat climate change must be, among other things, “sexy.” The English language statement was made last month ahead of a United Nations summit on the issue, but the use of that specific term by Koizumi, son of the popular former Prime Minister Junichiro Koizumi, has been widely scrutinized — prompting at least one lawmaker to demand an official explanation from the cabinet of current Prime Minister Shinzo Abe as to just what Koizumi meant. The answer arrived last week and it was … equivocal. A direct translation of the English word “sexy” into Japanese, the government said on Tuesday, was difficult to come by, in part because its meaning can “vary depending on the context.” While the demand for official explanation owed more to the ongoing hazing of the newly appointed Koizumi by opposition lawmakers than to a sincere quest for clarity, reporting on the government’s exegesis has taken on an exquisite air of parody. Meanwhile, when asked by Japanese reporters to explain himself, Koizumi’s response was deadpan: “The mere act of me explaining it,” The Japan Times quoted him as saying, “would not be sexy.” (Japan Times)


Michael Schulson is a contributing editor for Undark. His work has also been published by Aeon, NPR, Pacific Standard, Scientific American, Slate, and Wired, among other publications.