wildercb/PrivacyTextBooks1.2 · Datasets at Hugging Face

text stringlengths 0 118
resources into security.
The app stores disclaim responsibility. Although they often claim they
review apps before allowing them in the store, the app stores don’t want to
be held legally responsible. They simply want to review apps voluntarily
and avoid being held to a particular standard.
In Opperman v. Path, a group of plaintiffs sued Apple for
misrepresentation because it touted a strong commitment to privacy but
didn’t prevent apps from obtaining people’s contacts without consent. The
court threw out the case even though the court concluded that Apple
misrepresented the curating of its app store and the level of security of the
apps being sold there. Despite this misleading marketing, the court
concluded that it wasn’t extensive enough. Apple made “only a handful of
statements per year, contained in ‘buzz marketing’ materials, press releases,
statements in investor calls, and similar materials. These materials were not
widely disseminated to consumers.” In other words, Apple engaged in some
misrepresentation, but it didn’t engage in a tremendous amount of
misrepresentation. According to the court, a little wrongdoing is okay.31
The court totally ignored the fact that people generally assume that they
are safe in an app store. The platforms need not tout anything because they
can rely on the fact that people assume the products in their stores are
safe.32
Sometimes, the law flashes a burst of inspiration. A recent case is a
cause for celebration. In Oberdorf v. Amazon, a court held that Amazon
could be liable for a defective product sold by a third-party on Amazon’s
site. Oberdorf bought a collar for her dog. One day, the collar broke, and the
leash snapped back and hit Oberdorf in the face, permanently blinding her
in one eye. The seller of the dog collar, a third-party called “The Furry
Gang,” had all but vanished. The seller was unreachable, and its account on
Amazon had gone inactive.
Amazon argued that it wasn’t the seller of the collar because it “merely
provides an online marketplace for products sold by third-party vendors.”
The court noted that Amazon barely vets its third-party sellers:
Amazon’s Vice President of Marketing Business admitted that Amazon generally takes no
precautions to ensure that third-party vendors are in good standing under the laws of the country
in which their business is registered. In addition, Amazon had no vetting process in place to
ensure, for example, that third-party vendors were amenable to legal process.
The court held that Amazon should bear the responsibility for unsafe
products because Amazon “exerts substantial control over third-party
vendors.” The court noted that “Amazon is fully capable, in its sole
discretion, of removing unsafe products from its website.”33
Much of the time, the law doesn’t hold platforms accountable. Platforms
like Amazon haven’t yet been held responsible for selling insecure devices.
Platforms often disavow responsibility for the activity that occurs on them.
In the United States, Section 230 provides robust immunity for publishing
content by third-parties.34 Companies encourage activity on their platform,
make money from activity on their platform, and entice people to use their
platform and feel safe on it. They want all the profits and benefits without
the responsibility.
A common theme in the data ecosystem is that many actors want to reap
profits yet don’t want to pay the costs when something goes wrong with
security. The app makers don’t have sufficient incentive to make secure
apps. The platforms where the apps are sold don’t want to be their brother’s
keeper, though of course many make some effort to deny access to
fraudulent or insecure apps.35 Platforms have strong incentives to point to
the app makers when something goes wrong. Everyone makes money, and
consumers are lulled into the false belief that they are safe when, in fact,
they are not.
Amplifiers
Another group of actors amplify data security risks. Amplifiers sweep up
people’s data and store it all together, making an attractive treasure trove for
fraudsters. Despite the higher risks that storing all this data creates, many
amplifiers don’t provide better security than anyone else.
CONSUMER REPORTING AGENCIES
Equifax is one of the big three consumer reporting agencies with personal
data on hundreds of millions of people. In 2017, it had one of the largest
data breaches, facilitated by a few careless errors.
In March 2017, the United States Computer Emergency Readiness Team
(called US-CERT for short) sent Equifax an alert about Apache Struts,
which was open-source software that Equifax used.
The Apache Software Foundation had already issued a free new version
with the problem fixed. All Equifax had to do was to update to this new
version. Apache Struts was used by a number of employees for a number of
different applications, so multiple installations of the software had to be
updated.
When the Equifax security team received the alert, they quickly sent an
email around to about 400 employees informing them that if they were
responsible for an Apache Struts installation, then they should update the
software accordingly. But the mass email failed to include one employee
who maintained a dispute portal that used the Apache Struts software. This
installation wasn’t updated.
A week later, Equifax scanned the network for vulnerable versions of
Apache Struts, but the scan wasn’t properly configured and failed to detect
the Apache Struts installation on the dispute portal.
Four months later, Equifax’s security team noticed some suspicious
activity at the dispute portal. They later discovered that many attackers had
exploited the vulnerability in the portal to break into the network. To the
delight of the attackers, Equifax’s network wasn’t well segmented, so the
attackers were able to access a lot of information. They hit the jackpot when
they located administrative credentials which were stored in plain text. With
these credentials, they obtained even more access to Equifax’s network.
The hackers were even more gleeful when they were able to gain access
to more than 145 million records, including Social Security Numbers
(SSNs). Equifax’s policies required that personal data be encrypted, but the
SSNs were stored in plain text.36
This breach involves many of the themes we have discussed so far. The
vulnerability was in software that required a patch. A series of careless
human errors resulted in an installation of this software not being patched.
There were poor practices such as failing to encrypt data—even when
policies required it. The network wasn’t sufficiently segmented, giving the
intruders a lot more data to access. All these errors are careless human

resources into security.

The app stores disclaim responsibility. Although they often claim they

review apps before allowing them in the store, the app stores don’t want to

be held legally responsible. They simply want to review apps voluntarily

and avoid being held to a particular standard.

In Opperman v. Path, a group of plaintiffs sued Apple for

misrepresentation because it touted a strong commitment to privacy but

didn’t prevent apps from obtaining people’s contacts without consent. The

court threw out the case even though the court concluded that Apple

misrepresented the curating of its app store and the level of security of the

apps being sold there. Despite this misleading marketing, the court

concluded that it wasn’t extensive enough. Apple made “only a handful of

statements per year, contained in ‘buzz marketing’ materials, press releases,

statements in investor calls, and similar materials. These materials were not

widely disseminated to consumers.” In other words, Apple engaged in some

misrepresentation, but it didn’t engage in a tremendous amount of

misrepresentation. According to the court, a little wrongdoing is okay.31

The court totally ignored the fact that people generally assume that they

are safe in an app store. The platforms need not tout anything because they

can rely on the fact that people assume the products in their stores are

safe.32

Sometimes, the law flashes a burst of inspiration. A recent case is a

cause for celebration. In Oberdorf v. Amazon, a court held that Amazon

could be liable for a defective product sold by a third-party on Amazon’s

site. Oberdorf bought a collar for her dog. One day, the collar broke, and the

leash snapped back and hit Oberdorf in the face, permanently blinding her

in one eye. The seller of the dog collar, a third-party called “The Furry

Gang,” had all but vanished. The seller was unreachable, and its account on

Amazon had gone inactive.

Amazon argued that it wasn’t the seller of the collar because it “merely

provides an online marketplace for products sold by third-party vendors.”

The court noted that Amazon barely vets its third-party sellers:

Amazon’s Vice President of Marketing Business admitted that Amazon generally takes no

precautions to ensure that third-party vendors are in good standing under the laws of the country

in which their business is registered. In addition, Amazon had no vetting process in place to

ensure, for example, that third-party vendors were amenable to legal process.

The court held that Amazon should bear the responsibility for unsafe

products because Amazon “exerts substantial control over third-party

vendors.” The court noted that “Amazon is fully capable, in its sole

discretion, of removing unsafe products from its website.”33

Much of the time, the law doesn’t hold platforms accountable. Platforms

like Amazon haven’t yet been held responsible for selling insecure devices.

Platforms often disavow responsibility for the activity that occurs on them.

In the United States, Section 230 provides robust immunity for publishing

content by third-parties.34 Companies encourage activity on their platform,

make money from activity on their platform, and entice people to use their

platform and feel safe on it. They want all the profits and benefits without

the responsibility.

A common theme in the data ecosystem is that many actors want to reap

profits yet don’t want to pay the costs when something goes wrong with

security. The app makers don’t have sufficient incentive to make secure

apps. The platforms where the apps are sold don’t want to be their brother’s

keeper, though of course many make some effort to deny access to

fraudulent or insecure apps.35 Platforms have strong incentives to point to

the app makers when something goes wrong. Everyone makes money, and

consumers are lulled into the false belief that they are safe when, in fact,

they are not.

Amplifiers

Another group of actors amplify data security risks. Amplifiers sweep up

people’s data and store it all together, making an attractive treasure trove for

fraudsters. Despite the higher risks that storing all this data creates, many

amplifiers don’t provide better security than anyone else.

CONSUMER REPORTING AGENCIES

Equifax is one of the big three consumer reporting agencies with personal

data on hundreds of millions of people. In 2017, it had one of the largest

data breaches, facilitated by a few careless errors.

In March 2017, the United States Computer Emergency Readiness Team

(called US-CERT for short) sent Equifax an alert about Apache Struts,

which was open-source software that Equifax used.

The Apache Software Foundation had already issued a free new version

with the problem fixed. All Equifax had to do was to update to this new

version. Apache Struts was used by a number of employees for a number of

different applications, so multiple installations of the software had to be

updated.

When the Equifax security team received the alert, they quickly sent an

email around to about 400 employees informing them that if they were

responsible for an Apache Struts installation, then they should update the

software accordingly. But the mass email failed to include one employee

who maintained a dispute portal that used the Apache Struts software. This

installation wasn’t updated.

A week later, Equifax scanned the network for vulnerable versions of

Apache Struts, but the scan wasn’t properly configured and failed to detect

the Apache Struts installation on the dispute portal.

Four months later, Equifax’s security team noticed some suspicious

activity at the dispute portal. They later discovered that many attackers had

exploited the vulnerability in the portal to break into the network. To the

delight of the attackers, Equifax’s network wasn’t well segmented, so the

attackers were able to access a lot of information. They hit the jackpot when

they located administrative credentials which were stored in plain text. With

these credentials, they obtained even more access to Equifax’s network.

The hackers were even more gleeful when they were able to gain access

to more than 145 million records, including Social Security Numbers

(SSNs). Equifax’s policies required that personal data be encrypted, but the

SSNs were stored in plain text.36

This breach involves many of the themes we have discussed so far. The

vulnerability was in software that required a patch. A series of careless

human errors resulted in an installation of this software not being patched.

There were poor practices such as failing to encrypt data—even when

policies required it. The network wasn’t sufficiently segmented, giving the

intruders a lot more data to access. All these errors are careless human