Professional Documents
Culture Documents
NEWS
FEATURES
EVERYONE
J
ust as spring arrived last month Many academic publishers offer programs ing. Over the 6 months leading up to March,
in Iran, Meysam Rahimi sat down to help researchers in poor countries ac- Sci-Hub served up 28 million documents.
at his university computer and cess papers, but only one, called Share Link, More than 2.6 million download requests
immediately ran into a problem: seemed relevant to the papers that Rahimi came from Iran, 3.4 million from India, and
how to get the scientific papers he sought. It would require him to contact 4.4 million from China. The papers cover
needed. He had to write up a re- authors individually to get links to their every scientific topic, from obscure phys-
search proposal for his engineering work, and such links go dead 50 days after ics experiments published decades ago to
Ph.D. at Amirkabir University of a paper’s publication. The choice seemed the latest breakthroughs in biotechnology.
Technology in Tehran. His project clear: Either quit the Ph.D. or illegally ob- The publisher with the most requested Sci-
straddles both operations management and tain copies of the papers. So like millions of Hub articles? It is Elsevier by a long shot—
behavioral economics, so Rahimi had a lot of other researchers, he turned to Sci-Hub, the Sci-Hub provided half-a-million downloads
ground to cover. world’s largest pirate website for scholarly of Elsevier papers in one recent week.
But every time he found the abstract literature. Rahimi felt no guilt. As he sees it, These statistics are based on extensive
of a relevant paper, he hit a server log data supplied by
paywall. Although Amirk- Alexandra Elbakyan, the neuro-
abir is one of the top research scientist who created Sci-Hub in
universities in Iran, inter- “The numbers are just staggering. It suggests 2011 as a 22-year-old graduate
national sanctions and economic an almost complete failure to provide a path student in Kazakhstan (see bio,
woes have left it with poor ac- p. 511). I asked her for the data
cess to journals. To read a 2011 of access for these researchers.” because, in spite of the flurry of
paper in Applied Mathemat- Anonymous science publishing executive polarized opinion pieces, blog
ics and Computation, Rahimi posts, and tweets about Sci-
would have to pay the publisher, Elsevier, high-priced journals “may be slowing down Hub and what effect it has on research and
$28. A 2015 paper in Operations Research, the growth of science severely.” academic publishing, some of the most ba-
published by the U.S.-based company IN- The journal publishers take a very differ- sic questions remain unanswered: Who are
FORMS, would cost $30. ent view. “I’m all for universal access, but Sci-Hub’s users, where are they, and what are
He looked at his list of abstracts and did not theft!” tweeted Elsevier’s director of they reading?
the math. Purchasing the papers was go- universal access, Alicia Wise, on 14 March For someone denounced as a criminal
ing to cost $1000 this week alone—about as during a heated public debate over Sci-Hub. by powerful corporations and scholarly
much as his monthly living expenses—and “There are lots of legal ways to get access.” societies, Elbakyan was surprisingly forth-
he would probably need to read research pa- Wise’s tweet included a link to a list of 20 of coming and transparent. After establishing
pers at this rate for years to come. Rahimi the company’s access initiatives, including contact through an encrypted chat system,
was peeved. “Publishers give nothing to the Share Link. she worked with me over the course of several
authors, so why should they receive any- But in increasing numbers, researchers weeks to create a data set for public release:
thing more than a small amount for manag- around the world are turning to Sci-Hub, every download event over the 6-month pe-
ing the journal?” which hosts 50 million papers and count- riod starting 1 September 2015, including the
Published by AAAS
Corrected 13 May. Corrected 28 April 2016; see full text.
8. TUNISIA
487,720
4. RUSSIA
945,588
Moscow has 48%
of Russia’s total
downloads.
5. UNITED STATES
714,082 2. CHINA
Six the of top U.S. cities 10. MOROCCO 2,349,385
are from San Francisco 346,460 Beijing has 17%
Bay Area, Silicon Valley. 7. EGYPT of China’s total
9. INDONESIA
461,345
6. BRAZIL
562,056 1. IRAN 3. INDIA
REQUESTS 2,629,115 1,946,052
2,629,115 Downloaded eight times
1,000,000
more Sci-Hub articles than
its neighbor and rival Pakistan.
1
F. Nature Publishing H. 906,220 I. 880,343 3. Photosensitive feld emission study of SnS2 nanosheets
Group - 1,121,881 Journal of Vacuum Science & Technology B, 2015
2991
4. Grifths eDects and quantum critical points
in dirty superconductors without spin-rotation invariance:
B. Springer - 2,630,787 G. Royal Society of J. 377, 267 L. 302,525 N.
One-dimensional examples, Physical Review B, 2001
Chemisty - 927,238
2890
K. 358,786 O.
M. 206,294
5. Iron defciency: new insights into diagnosis
H. Informa UK (T&F) I. Wiley Blackwell (Blackwell Publishing) J. SAGE Publications K. JSTOR L. Oxford University Press M. AAAS and treatment, Hematology 2015, 2015
CREDITS: (DATA) SCI-HUB; (MAP) ADAPTED BY G. GRULLÓN/SCIENCE
250 K
200 K
150 K
100 K
50 K MISSING
DATA
0
SEPTEMBER OCTOBER NOVEMBER DECEMBER JANUARY FEBRUARY
Published by AAAS
Corrected 13 May. Corrected 28 April 2016; see full text.
Hub’s papers to make a local mirror of the reading in cities worldwide.) Mapping IP ton, D.C., has its science and technology
site, Elbakyan says. Rahimi, the engineering addresses to real-world locations can paint campus there, but Ashburn is also home to
student in Tehran, confirms this. “There are a false picture if people hide behind web Janelia Research Campus, the elite Howard
several Persian sites similar to Sci-Hub,” he proxies or anonymous routing services. But Hughes Medical Institute outpost, as well as
says. “So you should consider Iranian illegal according to Elbakyan, fewer than 3% of Sci- the servers of the Wikimedia Foundation,
[paper] downloads to be five to six times Hub users are using those. the headquarters of the online encyclope-
higher” than what Sci-Hub alone reveals. In the United States and Europe, Sci-Hub dia Wikipedia. Spokespeople for the latter
The geography of Sci-Hub usage gener- users concentrate where academic research- two say their employees are unlikely to ac-
ally looks like a map of scientific productiv- ers are working. Over the 6-month period, count for the traffic. The GWU press office
ity, but with some of the richer and poorer 74,000 download requests came from IP responded defensively, sending me to an on-
science-focused nations flipped. The smaller addresses in New York City, home to mul- line statement that the university recently
countries have stories of their own. Some- tiple universities and scientific institutions. issued about the impact of journal subscrip-
one in Nuuk, Greenland, is reading a paper There were 19,000 download requests from tion rate hikes on its library budget. “Schol-
about how best to provide cancer treatment Columbus, a city with less than a tenth of arly resources are not luxury goods,” it says.
to indigenous populations. Research goes New York’s population, and 68,000 from “But they are priced as though they were.”
on in Libya, even as a civil war rages there. East Lansing, Michigan, which has less than Several GWU students confessed to being
Someone in Benghazi is investigating a a hundredth. These are the homes of Ohio Sci-Hub fans. When she moved from Argen-
method for transmitting data between com- State University and Michigan State Univer- tina to the United States in 2014 to start
puters across an air gap. Far to the south in sity (MSU), respectively. her engineering Ph.D., Natalia Clementi
the oil-rich desert, someone near the town The numbers for Ashburn, Virginia, the says her access to some key journals within
B
eyond being the founder of Sci-Hub, possibility that such an interface could cumbersome: Post a request on Twitter
the world’s largest pirate site for one day translate the thought content to #IcanhazPDF with your email address.
academic papers, and risking arrest from one mind and upload it to another. Eventually, a generous researcher at some
as a result, Alexandra Elbakyan is But the work fell short of her dreams. university with access to the journal will
a typical science graduate student: “The lab activity was spiritless,” she says. send you the paper.
idealistic, hard-working, and relatively “There was no feeling of pursuing a What was needed, she decided, was
poor. In 1988, when Elbakyan was born higher goal.” a system that allowed that paper to be
in Kazakhstan, the Soviet Union was Elbakyan did find a community of like- shared—with absolutely everyone. She
just beginning to crumble. Books about minded researchers in transhumanism, had the computer skills—and contacts
dinosaurs and evolution fascinated her a lofty field that encompasses not just with other pirate websites—to make that
early on. “I also remember reading Soviet neuroscience and computer happen, and so Sci-Hub
science books that provided scientific technology but also philosophy was born (see main story, p.
explanations for miraculous events and even speculative fiction 508). Elbakyan sees the site
thought previously to be produced by about the future of humanity. as a natural extension of her
gods or magic.” She was hooked. She discovered a transhuman- dream of helping humans
At university in the Kazakh capital, she ism conference in the United share good ideas. “Journal
discovered a knack for computer hack- States and set her heart on paywalls are an example of
ing. It appealed to her because “unlike attending, but she struggled to something that works in the
higher programming languages that get a U.S. visa. She was rejected reverse direction,” she says,
are created by people and are volatile,” the first time and only barely “making communication
making and breaking computer security made it to the conference. With less open and efficient.”
systems requires a deeper knowledge of the remainder of her summer Alexandra Elbakyan, Running a pirate site and
mathematics and the primitive “assembly visa, she did a research intern- Sci-Hub founder. being sued for what is likely
language” that computers use to move ship at Georgia Institute of to be millions of dollars in
information. Technology in Atlanta. When she got back damages hasn’t stopped Elbakyan from
Like so many of Kazakhstan’s bright- to Kazakhstan, frustration with the bar- pursuing an academic career. Her neuro-
est, Elbakyan left the country to pursue riers that scientists face would soon lead science research is on hold, but she has
her dreams. First she worked in Moscow her to create Sci-Hub—an awe-inspiring enrolled in a history of science master’s
PHOTO: APNEET JOLLY/FLICKR
in computer security for a year, and then act of altruism or a massive criminal program at a “small private university”
she used the earnings to launch herself to enterprise, depending on whom you ask. in an undisclosed location. Appropriately
the University of Freiburg in Germany in Publisher paywalls are the bane of enough, her thesis focuses on scientific
2010, where she joined a brain-computer scientists and students in Kazakhstan, communication. “I perceive Sci-Hub as a
interface project. She was lured by the she says, and the existing solution was practical side of my research.” j
ing some specialty journals, she notes, but programs on it from an offline computer, journal access to the 240,000 research-
“most of them have no problem accessing Padilla had to walk the resulting data across ers of the University of California system.
big journals because the government pays campus on a thumb drive for analysis The authentication systems that university
the subscription at all the public universi- with Hart-Davidson. researchers must use to read subscription
ties around the country.” Yet Sci-Hub has drawbacks for text- journals from off campus, and even some-
Even for journals to which the university mining research, Hart-Davidson says. The times on campus with personal computers,
has access, Sci-Hub is becoming the go-to pirated papers are in unstructured PDF for- “are there to enforce publisher restrictions,”
resource, says Gil Forsyth, another GWU mat, which is hard for programs to parse. she says.
engineering Ph.D. student. “If I do a search But the bigger issue, he says, is that the data Will Sci-Hub push the industry toward an
on Google Scholar and there’s no imme- source is illegal. “How are you going to pub- open-access model, where reader authenti-
diate PDF link, I have to click through to lish your work?” Then again, having a mas- cation is unnecessary? That’s not clear, Har-
‘Check Access through GWU’ and then it’s sive private repository of papers does allow vard’s Suber says. Although Sci-Hub helps a
hit or miss,” he says. “If I put [the paper’s a researcher to rapidly test hypotheses be- great many researchers, he notes, it may also
title or DOI] into Sci-Hub, it will just work.” fore bothering with libraries at all. And it’s carry a “strategic cost” for the open-access
He says that Elsevier publishes the journals all just a click away. movement, because publishers may take
that he has had the most trouble accessing. advantage of “confusion” over the legality
The GWU library system “offers a docu- WHILE ELSEVIER WAGES a legal battle of open-access scholarship in general and
ment delivery system specifically for math, against Elbakyan and Sci-Hub, many in the clamp down. “Lawful open access forces pub-
physics, chemistry, and engineering faculty,” publishing industry see the fight as futile. lishers to adapt,” he says, whereas “unlawful
I was told by Maralee Csellar, the univer- “The numbers are just staggering,” one se- open access invites them to sue instead.”
Published by AAAS
Who's downloading pirated papers? Everyone
John Bohannon
RELATED http://science.sciencemag.org/content/sci/352/6285/511.full
CONTENT
http://science.sciencemag.org/content/sci/352/6285/497.full
PERMISSIONS http://www.sciencemag.org/help/reprints-and-permissions
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published by the American Association for the Advancement of
Science, 1200 New York Avenue NW, Washington, DC 20005. 2017 © The Authors, some rights reserved; exclusive
licensee American Association for the Advancement of Science. No claim to original U.S. Government Works. The title
Science is a registered trademark of AAAS.