wildercb/PrivacyTextBooks1.2 · Datasets at Hugging Face

text stringlengths 0 118
around clunky security procedures. When policymakers create rules that
don’t factor in people’s inevitable foibles and incentives to create
workarounds, they get unintended consequences. We need to think of ways
to better account for human behavior in designing security policy.
RETHINKING DESIGN RULES FOR TECHNOLOGIES
Although many experts realize that human fallibility is the bane of data
security, information technologies are often not designed with this fact in
mind. By designing, we mean creating something according to a plan.
Design is a key aspect of good data security policy.49 Design is everywhere,
design is power, and design is political.50 Nearly everything with
technology results from a design choice, from the interface we see when we
use software, to the buttons we push on devices, to the initial settings on
platforms. The way humans interact with technology is at the center of so
many security vulnerabilities, and design is all about structuring these
interactions.
In writing about the pitiful security in most modern technological
devices and systems, Bruce Schneier suggests that “Security needs to be
engineered into every system, and every component of every system, from
the beginning and throughout the development process.”51 We agree. The
principle that Schneier is recommending is known as “security by design.”
The law is several steps behind. First, only a few data security laws even
try to regulate design. Most notably, Article 25 of the GDPR requires “data
protection by design and by default.”52 Although the GDPR isn’t very
specific on what these baselines should be, at least it is a start. Beyond the
GDPR, few other laws say anything about design. The law must change
course and start regulating design, as doing so is crucial for strong data
security.53
Second, the law not only must regulate design, but also must push for
design that accounts for the human element. Designs that are oblivious to
human behavior will fail. Security often focuses on keeping attackers out,
rather than mitigating the errors from people who are already properly in.
All design choices have one of two effects (and often both): They can
signal information to people or make particular behavior easier or harder
by imposing or reducing transaction costs—the expenditures of time, labor,
and resources that are necessary to complete a task. Signals and transaction
costs shape our mental models of how technologies work and form the
constraints that guide our behavior in particular ways.54 Design decisions
make certain realities more or less likely. Companies and hackers know that
the power to design is the power to exert control over people. This is
because people react to signals and constraints implemented through design
in predicable ways. People with the right know-how can build or leverage
the design of consumer technologies to encourage desired user behavior.55
A concern with regulating design is that it might be too paternalistic. We
don’t want policymakers not versed in technology to make decisions far
beyond their expertise. The regulation of design, however, need not focus
on the specific details of technologies; policymakers do not need to become
backseat engineers. Instead, there are some more general design
requirements that the law can impose that will not be overly meddlesome
but will be quite effective in designing more securely. Below, we set forth a
few proposals for the types of security design requirements that follow our
overarching principle, which is to design with the human element in mind.
Changing the Default Settings
Humans are prone to inertia. Most users will not take action, and design
must reflect this fact. We can beg and plead with people to do things, but
this nagging will fail to work consistently with many people. Bruce
Schneier wisely recommends that all “devices need to be secure without
much intervention by users.”56 The less work people need to do for security,
the better.
One way design can have a significant effect is with default settings—
the preselected options within software programs and devices. Default
settings have a major impact on human behavior. Many programs, apps, and
devices have initial settings that are quite insecure. For example, the default
settings for many social media sites are set to maximize sharing. According
to a 2015 study from Kaspersky Lab, 28 percent of social media users
neglect privacy settings and leave all their posts, photos, and videos
accessible to the public.57 Not only is this setting bad for privacy, but it is
also bad for security as it can lead to people exposing personal information
in ways that might compromise their overall data security. Recall how
Christopher Chaney was able to guess the password recovery questions of
celebrities: he found their information online.
Certain apps and sites have unexpected defaults that can be very clunky
to change. For example, at Venmo, a payment service, user financial
transactions are public by default.58 The process for a user to limit access to
all their transactions is not very intuitive. The “Default Audience” setting
only affects the visibility of the charges and payments initiated by the user.
The transactions initiated by friends who have not changed the default
privacy setting will still be shared publicly. To obscure all transactions, a
user must hunt down and change the inconspicuous “Transactions Involving
You” setting.59
The law could require or encourage default settings limiting the sharing
of personal information online. If people want to share more, they could
certainly change the settings, but many people haven’t thought enough
about the privacy and security consequences of the settings. As we
discussed in the previous chapter, rules that strengthen privacy often also
strengthen security, and a rule about the default settings is one that is often
in the privacy category but can have benefits for security.
The law could also require or encourage default settings for the use of
two-factor authentication for certain services or the option to automatically
update software to more secure versions. If these options are not selected by
default (or if the software doesn’t nudge users towards using them), then
users are less likely to take advantage of them.
Another way that the law can help strengthen security is to address
default passwords. One of the most commonly exploited vulnerabilities in
devices is the default password.60 Many devices have a simple password by
default, and it is sometimes hard-coded in. Many users never change this
password, making it easy for hackers to break into devices. This is how the
previously discussed DDoS attack on Krebs was perpetrated.
One solution would be to require manufacturers of devices to require
that users change default passwords when they start using the device. This
rule might seem like a small requirement, but it can make an enormous

around clunky security procedures. When policymakers create rules that

don’t factor in people’s inevitable foibles and incentives to create

workarounds, they get unintended consequences. We need to think of ways

to better account for human behavior in designing security policy.

RETHINKING DESIGN RULES FOR TECHNOLOGIES

Although many experts realize that human fallibility is the bane of data

security, information technologies are often not designed with this fact in

mind. By designing, we mean creating something according to a plan.

Design is a key aspect of good data security policy.49 Design is everywhere,

design is power, and design is political.50 Nearly everything with

technology results from a design choice, from the interface we see when we

use software, to the buttons we push on devices, to the initial settings on

platforms. The way humans interact with technology is at the center of so

many security vulnerabilities, and design is all about structuring these

interactions.

In writing about the pitiful security in most modern technological

devices and systems, Bruce Schneier suggests that “Security needs to be

engineered into every system, and every component of every system, from

the beginning and throughout the development process.”51 We agree. The

principle that Schneier is recommending is known as “security by design.”

The law is several steps behind. First, only a few data security laws even

try to regulate design. Most notably, Article 25 of the GDPR requires “data

protection by design and by default.”52 Although the GDPR isn’t very

specific on what these baselines should be, at least it is a start. Beyond the

GDPR, few other laws say anything about design. The law must change

course and start regulating design, as doing so is crucial for strong data

security.53

Second, the law not only must regulate design, but also must push for

design that accounts for the human element. Designs that are oblivious to

human behavior will fail. Security often focuses on keeping attackers out,

rather than mitigating the errors from people who are already properly in.

All design choices have one of two effects (and often both): They can

signal information to people or make particular behavior easier or harder

by imposing or reducing transaction costs—the expenditures of time, labor,

and resources that are necessary to complete a task. Signals and transaction

costs shape our mental models of how technologies work and form the

constraints that guide our behavior in particular ways.54 Design decisions

make certain realities more or less likely. Companies and hackers know that

the power to design is the power to exert control over people. This is

because people react to signals and constraints implemented through design

in predicable ways. People with the right know-how can build or leverage

the design of consumer technologies to encourage desired user behavior.55

A concern with regulating design is that it might be too paternalistic. We

don’t want policymakers not versed in technology to make decisions far

beyond their expertise. The regulation of design, however, need not focus

on the specific details of technologies; policymakers do not need to become

backseat engineers. Instead, there are some more general design

requirements that the law can impose that will not be overly meddlesome

but will be quite effective in designing more securely. Below, we set forth a

few proposals for the types of security design requirements that follow our

overarching principle, which is to design with the human element in mind.

Changing the Default Settings

Humans are prone to inertia. Most users will not take action, and design

must reflect this fact. We can beg and plead with people to do things, but

this nagging will fail to work consistently with many people. Bruce

Schneier wisely recommends that all “devices need to be secure without

much intervention by users.”56 The less work people need to do for security,

the better.

One way design can have a significant effect is with default settings—

the preselected options within software programs and devices. Default

settings have a major impact on human behavior. Many programs, apps, and

devices have initial settings that are quite insecure. For example, the default

settings for many social media sites are set to maximize sharing. According

to a 2015 study from Kaspersky Lab, 28 percent of social media users

neglect privacy settings and leave all their posts, photos, and videos

accessible to the public.57 Not only is this setting bad for privacy, but it is

also bad for security as it can lead to people exposing personal information

in ways that might compromise their overall data security. Recall how

Christopher Chaney was able to guess the password recovery questions of

celebrities: he found their information online.

Certain apps and sites have unexpected defaults that can be very clunky

to change. For example, at Venmo, a payment service, user financial

transactions are public by default.58 The process for a user to limit access to

all their transactions is not very intuitive. The “Default Audience” setting

only affects the visibility of the charges and payments initiated by the user.

The transactions initiated by friends who have not changed the default

privacy setting will still be shared publicly. To obscure all transactions, a

user must hunt down and change the inconspicuous “Transactions Involving

You” setting.59

The law could require or encourage default settings limiting the sharing

of personal information online. If people want to share more, they could

certainly change the settings, but many people haven’t thought enough

about the privacy and security consequences of the settings. As we

discussed in the previous chapter, rules that strengthen privacy often also

strengthen security, and a rule about the default settings is one that is often

in the privacy category but can have benefits for security.

The law could also require or encourage default settings for the use of

two-factor authentication for certain services or the option to automatically

update software to more secure versions. If these options are not selected by

default (or if the software doesn’t nudge users towards using them), then

users are less likely to take advantage of them.

Another way that the law can help strengthen security is to address

default passwords. One of the most commonly exploited vulnerabilities in

devices is the default password.60 Many devices have a simple password by

default, and it is sometimes hard-coded in. Many users never change this

password, making it easy for hackers to break into devices. This is how the

previously discussed DDoS attack on Krebs was perpetrated.

One solution would be to require manufacturers of devices to require

that users change default passwords when they start using the device. This

rule might seem like a small requirement, but it can make an enormous