7 December 2019 | New Scientist | 9MENTAL health websites are
sharing user data with advertisers,
including the results of tests for
depression. This means that
people seeking information or
help for mental health conditions
can be targeted with adverts while
they may be vulnerable.
Eliot Bendinelli at Privacy
International in London and his
colleagues looked at 136 of the
most popular websites in the UK,
France and Germany that provide
resources and information about
mental health conditions. The
researchers found that 76 per cent
of the websites contained third-
party marketing trackers. These
collect information about a user
and can track them as they browse
other sites. This can be combined
into a detailed profile.
Many of the pages had trackers
from Google, Facebook and
Amazon and shared information
with data brokers – firms that
aggregate information and sell
individual profiles to other
organisations – and advertising
companies.
“It’s currently almost
impossible to seek information
and help about depression
without advertisers knowing,”
says Frederike Kaltheuner at the
Mozilla Foundation in London,
who is part of the team. “Knowing
who is depressed and when allows
advertisers to target people when
they are at their most vulnerable.
Feeling low today? Here are some
diet pills.”
Advertisers target users based
on their personal data, such as
their IP address and location, and
the site they visit, says Bendinelli.
Some sites also have real-time
bidding, where information,
including a page’s content and
URL, is used to instantaneously
show a relevant ad on the page.
Several websites with
questionnaires about depressionstored users’ answers and shared
them with third parties.
When the researchers first
analysed the UK’s National Health
Service website in September,
they found that a mood self-
assessment quiz shared individual
answers, test scores and the test
URL with Adobe for analysis
purposes. However, the NHS
website has since updated its
privacy policy so users now need
to manually opt in to be tracked.
The team also found that
doctissimo.fr, one of France’s
most popular health websites,
sent unique user identifiers and
answers to a data collection firm.
“Doctissimo basically sends all of
your answers to the test to a third
party that is named nowhere on
their website, nowhere in their
privacy policy,” says Bendinelli.
“That’s a perfect example of
abusing people’s confidence.”
The research was presented
this week at the Black Hat Europe
event in London.Under the EU’s General Data
Protection Regulation, websites
and apps are required to obtain
consent before tracking users.
Data relating to physical or
mental health is considered a
“special category” and can only
be processed with explicit consent
or for relevant other reasons.
Doctissimo.fr didn’t respond
to New Scientist’s requests for
comment. Google and Amazon
say they don’t use sensitive data
for advertising and that they
prohibit advertisers from doing so
as well. Facebook says it requires
website owners to be clear about
the information they are sharing
with the firm. ❚A ROMAN shipwreck found off
the coast of Cephalonia Island,
Greece, may reveal clues about
the beginning of the Roman
Empire, such as how people
conducted trade and the
wider economy.
George Ferentinos at the
University of Patras and his
colleagues explored the sea
floor around Cephalonia
Island in the Ionian Sea
using sonar imaging.
They discovered three wrecks
from the second world war –
two ships and one plane – that
are almost intact, as well as a
large ancient Roman vessel that
they dubbed the Fiscardo, after
the present-day fishing port
near where it was discovered.
The top of the 2000-year-old
Roman wreck is full of jug-like
containers called amphorae.
Their distinctive shapes allowed
the team to date the wreck to
between the 1st century BC
and the 1st century AD (Journal
of Archaeological Science,
doi.org/ggdqcr).
At 34 metres long and
13 metres wide, the Fiscardo
is one of the largest four
shipwrecks from this periodfound in the Mediterranean
Sea, and the largest in the
east Mediterranean Sea,
says Ferentinos.
“It’s half-buried in the
sediment, so we have high
expectations that if we go to an
excavation in the future, we will
find part or the whole wooden
hull,” he says. This could tell
archaeologists when and where
the ship was made, where the
material came from and how
it was repaired.
Judging by the ship’s size, it
was probably carrying around
6000 amphorae at the time
it sank, he says. These were
typically used to transport wine,
olive oil, grains and other goods.
This is a monumental number,
says Craig Barker at the
University of Sydney, Australia.
The size, shape and design
of these containers vary by
location, so could reveal where
the ship was sailing to and
from – and DNA and chemical
analyses may even reveal what
was inside them. ❚Archaeology Data privacyRuby Prosser Scully Donna LuGE
OR
GE
FE
RE
NT
INO
S^ E
T^ A
L.
Roman shipwreck
full of 2000-year-
old jugs discovered
“It’s currently almost
impossible to seek help
about depression without
advertisers knowing”Mental health sites share
user data with ad firms
The Roman ship may have
held 6000 jug-like containers
called amphorae