SYSTEMATIC EVALUATION AND USERSTUDY OF PRIVACY OF DEFAULT APPS IN APPLE SMOBILE ECOSYSTEM Amel Bourdoucen

2025-05-03 0 0 1017.22KB 23 页 10玖币

侵权投诉

SYSTEMATIC EVALUATION AND USER STUDY OF PRIVACY OF

DEFAULT APPS IN APPLE’SMOBILE ECOSYSTEM

Amel Bourdoucen

Aalto University

Janne Lindqvist

Aalto University

ABSTRACT

Users need to conﬁgure default apps when they ﬁrst start using their devices. The privacy conﬁg-

urations of the default apps do not always match what users think they have initially enabled. We

ﬁrst systematically evaluated the privacy conﬁgurations of default apps. We discovered serious issues

with the documentation of the default apps. Based on these ﬁndings, we explored users’ experiences

with an interview study (N=15). Our ﬁndings from both studies show that: the instructions of setting

privacy conﬁgurations of default apps are vague and lack required steps; users were unable to disable

default apps from accessing their personal information; users assumed they were being tracked by

some default apps; default apps may cause tensions in family relationships because of information

sharing. Our results illuminate on the privacy and security implications of conﬁguring the privacy of

default apps and how users perceive and understand the mobile ecosystem.

Keywords Privacy ·Mobile Devices ·Apps ·Ecosystems

1 Introduction

Users are not in control of their privacy preferences in complex mobile device and cloud ecosystems. The complexity

of these systems enable helpful features, but with the cost of privacy. Often, many of these features are enabled by

default when a user ﬁrst starts using the device(s). When users purchase new devices, they are often presented with

many features to enable or disable. Unfortunately, our work shows that the privacy implications of these features are

not understood and the settings are non-trivial to conﬁgure correctly.

Studies have discovered several challenges that users face in understanding how their data is handled when using

mobile apps [

Liccardi et al.(2014)

Palmerino(2018)

Liu et al.(2014)

Van Kleek et al.(2017)

Thompson et al.(2013)

Users are often presented with many permissions requesting approval of access to their personal data. More-

over, prior work suggests that users do not always fully understand the implications of enabling privacy conﬁg-

urations [

Frik et al.(2022)

Breitinger et al.(2019)

]. This confusion is often caused by the architecture of privacy

conﬁgurations [

Egelman et al.(2013)

]. Users are left confused [

Tan et al.(2014)

], surprised [

Balebako et al.(2013)

Shklovski et al.(2014)

], anxious [

Coopamootoo et al.(2022)

] and sometimes frustrated [

Shklovski et al.(2014)

] when

learning what actually happens to their data.

In this paper, we study default apps in Apple’s iOS and macOS

. MacOS and iOS include setup wizards that guide users

to selecting preferences for these default apps when a new device is started. As a concrete example, in the Setup Wizard,

one of the features that users can setup is Siri. Users are offered the choice to either Continue (to set up Siri) or Set Up

Later in Settings. Users can select Set Up Later in Settings, which implies that Siri is disabled until users set it up later.

This is despite selecting the option to Set Up Later in Settings. To systematically study these issues, we focused on the

following research questions:

RQ1: What privacy conﬁgurations are available to control default apps?

RQ2: How can users control default apps?

RQ3: How do users understand privacy conﬁgurations and their privacy and security implications?

1This work is an independent publication and has not been authorized, sponsored, or otherwise approved by Apple Inc.

arXiv:2210.04569v1 [cs.HC] 10 Oct 2022

RQ4: How does setting up default features impact privacy of users?

To answer these research questions, we conducted two studies; in the ﬁrst study we selected eight default apps and

features on iOS and macOS: Safari (Web browsing), Family Sharing (Shared Access), Siri (Virtual Assistant), iMessage

(Messaging), Facetime (Video Calls), Location Services and TouchID (Fingerprint). We analyzed Apple’s ofﬁcial

documentation and found them to seriously lack required details for conﬁguring privacy. Due to this, we conducted a

comprehensive system evaluation to understand the privacy conﬁgurations for these default apps. We mapped the routes

to disable features of these apps. Based on this ﬁrst study, we conducted a follow-up user study to investigate users’

understanding of privacy conﬁgurations. We compared the users answers to our ﬁndings from Study 1.

We make the following major contributions:

In our ﬁrst study, we thoroughly collected and investigated the privacy conﬁgurations of eight default apps of

the Apple mobile ecosystem. We present what user data is collected by each default app as well as paths to

disable different privacy conﬁgurations (presented in Section 3). We observed several serious issues associated

with conﬁguring default apps.

In the second study, we conducted a user study to explore users’ perceptions of using default apps and

conﬁguring their privacy and security settings. Our ﬁndings show that users had limited understanding of what

data was being collected and how it was shared (presented in Section 6).

Based on the two studies, we made the following major ﬁndings: participants were surprised to know some

privacy conﬁgurations were turned on by default; participants were confused about what happened to their

information; some participants assumed they were being tracked by some apps; tracking can introduce trust

issues in families and relationships; participants were aware that disabling data sharing does not guarantee

data is not shared anymore; participants wanted to know what happens to their information; participants did

not know how to disable data sharing on default apps.

2 Related Work

We ﬁrst discuss the most closely related prior work and proceed to discuss studies that explored users general

understanding of privacy in apps. We then describe challenges that users face when setting privacy conﬁgurations to

gain control over their personal data. Most closely related to our work are the following three studies:

First, a study [

Frik et al.(2022)

] explored the relationship between socio-economic factors and users’ choices of security

and privacy settings. The study surveyed users about the security and privacy settings of their mobile device’s operating

system (Android and iOS) and pre-installed browsers (Safari and Chrome). These settings included passcode, face

unlock, automatic updates, and password re-use. Participants were also asked about privacy and security risks the are

concerned of. The study found differences between socio-demographic groups on using security and privacy settings.

For example, older adults were found to worry less about online risks. The study also found that many users were not

aware of the function of settings used in the study but were willing to change them in the future.

Second, another study [

Gamba et al.(2020)

] explored pre-installed apps on the Android platform for their app packages,

certiﬁcates, and third-party libraries. This work revealed various actors involved in the development of pre-installed

apps. Potentially harmful behavior was detected from pre-installed apps related to personal data collection. The study

validated which app collected which set of personal information by collecting ﬁrmware and trafﬁc information from

users.

Third, another study [

Ramokapane et al.(2019)

] explored users’ awareness of some features on both Android and iOS.

These features were 1) location, 2) ads tracking and 3) usage and diagnostics (this is called Analytics Data Sharing in

iOS). The study asked participants to conduct four tasks on either a iOS or Android device based on the participant’s

experience and preference. These tasks were: Disable location services, Restrict App from using a default feature (iOS:

Camera, Android: disable Google app having access to Calendar and Location services), Disable/Limit ad tracking,

Restrict Usage and Diagnostic Report (Android only) or Disable analytical data sharing (iOS only). The participants

were given devices owned by the authors that were reset so that these tasks to be completed. The authors enabled the

settings of the features and asked users to locate and disable them. The participants performed a cognitive walk trough

and think aloud protocols while performing these tasks. The ﬁndings showed that participants were not able to easily

locate the settings of the features used in the study. Participants attributed these challenges to hidden controls and

complex app requests. To overcome these challenges, participants used quick ﬁxes or coping strategies such as skipping

privacy conﬁguration requests or searching the internet for quick solutions.

In contrast to the above three prior work, we evaluated the privacy conﬁgurations available in the Apple iOS and macOS

systems and speciﬁcally probed into users’ understanding of privacy in the Apple ecosystem. Towards this end, we

systematically tested setting up both iOS and macOS devices and documented all the steps required to disable settings.

We used this information as the basis of our qualitative interviews towards the end of probing to users’ understandings

and experiences with these privacy conﬁgurations and devices. We asked our participants to use their own devices rather

than test devices provided by us. We wanted our participants to perform the tasks with i) familiar devices and ii) devices

that contain their own privacy conﬁgurations that they have setup. Our results reﬂected users’ actual experiences with

privacy settings they have conﬁgured on their own rather than if we pre-conﬁgured it for them. Our work contributes by

revealing the difﬁculties on properly conﬁgure privacy for features that are enabled by default, and the surprises and

tensions that users experiences because of these settings.

Next, we will discuss a body of research that explored privacy of users when using mobile apps and conﬁguring privacy

settings.

User Privacy in Mobile Apps

Prior work has largely explored app permissions in mobile devices. Studies have

been motivated to explore users conﬁguring permissions due to the reported difﬁculties when doing so. For several

years, researchers have focused on permissions of apps in an attempt to explore ways to improve users’ understand-

ing and expectations when setting privacy conﬁgurations of apps. The focus on app permissions was motivated

through the many studies that demonstrated users’ difﬁculties in understanding privacy conﬁgurations of mobile apps

[Balebako et al.(2013), Ramokapane et al.(2019), Tan et al.(2014), Liccardi et al.(2014)].

A study on 308 Android users revealed that only 17% of users were attentive to the permissions that were prompted

during app installations, therefore indicated that permission warnings are not sufﬁcient to make informed security

decisions [

Felt et al.(2011)

]. A recent study in 2021 on 4,636 Android users has also conﬁrmed that information

provided by the current system is not enough for users to make informed decisions on their privacy [

Shen et al.(2021)

Other studies also showed that often users either ignore or accept permissions without reading the details properly

[Felt et al.(2012a), Felt et al.(2012b), Ramokapane et al.(2019)].

Researchers highlighted factors that inﬂuence the misunderstandings that users have of privacy conﬁgurations. Several

factors make it harder for users to know what happens to this personal information when agreeing to permissions or

conﬁgurations; for example, unclear Privacy Policies, and lack of transparency about data practices. Earlier studies have

investigated ways to improve Privacy Policies for better delivery for users [

Kelley et al.(2009)

]. However, recent studies

have reported that the unclear nature of Privacy Policies of apps still contribute to the difﬁculty users have to grasp

what happens to their personal data [

Alohaly and Takabi(2016)

Coen et al.(2016)

Kelly et al.(2012)

]. As a result of

the unclear nature of Privacy Policies, users rarely follow Privacy Policy links to read what part of their information is

disclosed [

Coen et al.(2016)

]. Another factor that has been suggested to contribute to the difﬁculty in understanding

what happens to personal data in apps is transparency [

Liccardi et al.(2014)

Van Kleek et al.(2017)

]. A study found

that providing more transparency to users about what occurs to their personal data can make users more conﬁdent

in their app use [

Van Kleek et al.(2017)

]. To help users better understand how their personal data is handled, recent

literature work explored several solutions to help make more informed decisions [

Liccardi et al.(2014)

Liu et al.(2014)

Palmerino(2018)

Thompson et al.(2013)

Van Kleek et al.(2017)

]. For instance, a study deployed machine learning to

offer a prospect of mitigating the burden of increased privacy decisions [

Smullen et al.(2020)

]. Another study proposed

a prototype that adjusted privileges given to apps on iOS as well as the ability to replace real data with mock data

[

Lutaaya(2018)

]. Another study analysed settings of 4.8 million smartphone users and demonstrated a number of

proﬁles that aim to simplify the decisions mobile users have to make about their privacy [Liu et al.(2014)].

Impacts of setting privacy conﬁgurations

A second line of research focused on understanding user’s concerns

when it comes to setting privacy preferences [

Balebako et al.(2013)

Ramokapane et al.(2019)

Shen et al.(2021)

Wijesekera et al.(2018)

]. Research has found that users often have misconceptions about the data sharing that oc-

curs on smartphone apps [

Balebako et al.(2013)

Frik et al.(2022)

Ramokapane et al.(2019)

Tan et al.(2014)

]. Mis-

conceptions about the handling of personal data can create challenges to users. Studies suggested that users

may feel uncomfortable or confused when learning about what occurs to their data [

Tan et al.(2014)

]. For ex-

ample, studies suggested that users are often surprised when asked to share their personal data collected by apps

[

Balebako et al.(2013)

Shklovski et al.(2014)

] for example, users understood that data was used for purposes such as

marketing but were surprised by the scope of data sharing, frequency and destination [Balebako et al.(2013)].

Users can also experience other emotions such as confusion about certain personal data that is requested, sometimes dis-

may or outrage [

Shklovski et al.(2014)

]. Studies that focused on tracking of users by apps [

Coopamootoo et al.(2022)

Ur et al.(2012)

McDonald and Cranor(2010)

], suggest that users can feel negatively under the perception of being

tracked. These feelings can include: anger, distrust and anxiety. Often users would feel accepting of the fact that

they are tracked, under certain conditions [

Coopamootoo et al.(2022)

]. The reactions users had upon learning about

how their data is handled can be denoted to the insufﬁcient information on mobile apps to help users make informed

decisions [

Ramokapane et al.(2019)

]. Insufﬁcient information provided about data handling may also lead users to

think that these permissions are required for apps to run [Smullen et al.(2020)] which inﬂuences users to accept them.

Summary

Prior work has focused extensively on privacy settings of mobile devices. An example of this

is users’ attitudes towards permissions of different apps on mobile devices [

Felt et al.(2012a)

Felt et al.(2012b)

Ramokapane et al.(2019)

], tracking of apps [

Coopamootoo et al.(2022)

McDonald and Cranor(2010)

Ur et al.(2012)

]

and privacy conﬁgurations of apps [

Balebako et al.(2013)

Frik et al.(2022)

Ramokapane et al.(2019)

Tan et al.(2014)

]. Although prior work focused extensively on users’ attitudes when setting permissions of

apps on mobile devices, little research (presented above) has focused on users’ experiences when setting up privacy

conﬁgurations of default apps.

3 Study I: Mobile Ecosystem Evaluation

Table 1: Privacy conﬁgurations of default apps. These privacy conﬁgurations may lead to personal data of users being

transferred outside the device as shown in the last column. For the full table, refer to Appendix A.1

Default App N Steps Privacy Conﬁgurations May transfer to Cloud or Vendor’s Servers

Safari N>12 IP Address Yes

Private Browsing No

Web Page Translation Translation locally, other data may leave device

iCloud Syncing Yes

Preload Top Hit in Safari Information not provided by vendor

Sending Information to Apple Yes

History and Website Data Yes

Siri N>9 Ask Siri Yes

Integrated apps Yes

Siri and Dictation Yes

Siri Personalisation Yes

iCloud Syncing Yes

Location Services Yes

Request History Yes

Facetime N>7 Enable Facetime Actual calls No, Otherwise Yes

Caller ID Yes

SharePlay Information not provided by vendor

Speaking Information not provided by vendor

FaceTime Live Photos Information not provided by vendor

Blocked Contacts Information not provided by vendor

In our ﬁrst study, we analyze the main eight features of the Apple iCloud mobile ecosystem. Today, we have few

mobile ecosystems that are popular, for example, Apple’s [

Apple(2022a)

], Google’s [

Google(2022)

] and Huawei’s

[

Huawei(2022)

]. As the target of our work, we choose Apple’s ecosystem for the following reasons: (1) Apple’s devices

are popularly purchased worldwide. In 2021, Apple reported a 65.6 billion USD revenue in iPhones only in the ﬁrst

quarter of the same year [

Hartmans(2021)

]. (2) Apple’s mobile ecosystem is uniquely cohesive and its integrated model

provides a quality experience for users, alongside a stated emphasis on privacy and security.

3.1 Mobile Ecosystem Structure

A mobile ecosystem consists of a set of units (devices) interacting with each other through exchange of information,

resources and artifacts [

Campbell and Ahmed(2010)

]. For instance, Apple’s mobile ecosystem consists of: iCloud

(Apple’s Cloud system) which brings together devices such as: iPad, iMac, Macbook, and iPhone. Default apps are

central to users’ participation in the ecosystem. The apps manufactured by Apple are referred to in this work by Default

apps. For instance, a popular app Safari is used as a browser for Apple devices. Information stored in Safari such as

Bookmarks are exchanged between devices that are connected to the same iCloud account.

Table 2: Users’ personal data collected from default apps. Users are not able disable the collection of some of the data

below for an app to function [Apple(2021)].

Default app Users’ personal data collected (not limited to list below)

Safari

IP address, sites you visit; open tabs, tab groups, AutoFill information, Bookmarks,

Reading List and History, attribution reports, payment method information.

Siri

Contact names, nicknames and relationships, Music and Podcasts, Names of your and

your Family Sharing members’ devices, Accessories, Homes, Scenes and Members

of Shared Home in Home app, Labels for Items (e.g., people’s names in Photos and

Alarms), Name of apps and shortcuts.

iMessage Articles, TV shows, Music and Photos.

Facetime

Facetime Calls (e.g., who was invited to call, device network conﬁgurations), Apps

using Facetime, Phone numbers, email addresses associated with account.

Family Sharing

Apple Watch serial number, cellular hardware identiﬁers, family member’s health,

location and contact data, view logs and screenshots from Apple Watch.

Touch ID 360-degree orientation ﬁngerprint data, passcode.

Location

Location data, Location Search Query, Geo-tagged locations of nearby WiFi hotspots,

GPS data, travel speed, barometric pressure, places you recently been, IP.

Find My

Participation in Find My network, device location, information about device, informa-

tion about account.

3.2 Method of Study I

We evaluated the system for the following parameters: deﬁning mobile ecosystems’ structure, privacy conﬁgurations of

default apps, number of privacy conﬁgurations to disable within an app and types of personal data collected by these

default apps and whether the personal data is transferred outside the device. We then analysed in depth Apple’s eight

default apps: Safari, Family Sharing, Find My, iMessage, Facetime, Siri, Location Services and TouchID. These apps

are linked to simple conﬁguration options presented to the user when the user ﬁrst time starts using macOS or iOS

device.

We had the following three major tasks in the analysis:

1. Analysis of Ofﬁcial Sources: We ﬁrst read the ofﬁcial sources provided by Apple [Apple(2021)]. There were

several challenges to perform this analysis;

Closed Ecosystem: Apple’s ecosystem is closed; meaning that

some of the speciﬁcations about the processing of personal data by default apps are not disclosed. Examples of

non-disclosure includes ambiguous phrasing such as “subsets of data stored” [

Apple(2021)

] without indicating

what the subset of data includes, how is it processed and for how long it is retained.

Scattered information:

When reading Apple’s Privacy Policy for the steps to disable a feature, we discovered that the controls of some

apps are described under other apps. For example, Siri has a speciﬁc section where its controls are described.

However, Siri can also be found under Safari Search’s section as well as Dictation. In summary, the public

Privacy Policies are long and do not contain easily accessible information how to precisely control privacy

conﬁgurations.

System Navigation: To be able to present the steps required to setup a device, we captured the setup and usage

processes on sample iOS and macOS devices. We ﬁrst did factory resets to our test devices before following

steps of the setup wizard in all possible combination of scenarios and sequence of steps. We repeated the latter

process every time we followed a different sequence of steps to start fresh).

Mapping of Privacy Conﬁgurations: To obtain on how many steps are required to disable each app, we mapped

the privacy conﬁgurations responsible for handling personal data for each app and noted the pathways to each

privacy control. We emphasized what privacy control was included in Apple’s sources and what was not

included. We also present the personal data collected from users’ by each app.

3.3 Results of Study I

We present a brief summary results of Study I below. Further details are given in the appendices.

文档加载中……请稍候！
如果长时间未打开，您也可以点击刷新试试。

下载文档到电脑，查找使用更方便

10 玖币 0人已下载

立即下载

摘要：

SYSTEMATICEVALUATIONANDUSERSTUDYOFPRIVACYOFDEFAULTAPPSINAPPLE'SMOBILEECOSYSTEMAmelBourdoucenAaltoUniversityJanneLindqvistAaltoUniversityABSTRACTUsersneedtoconguredefaultappswhentheyrststartusingtheirdevices.Theprivacycong-urationsofthedefaultappsdonotalwaysmatchwhatusersthinktheyhaveinitiallyenab...

展开>> 收起<<

SYSTEMATIC EVALUATION AND USERSTUDY OF PRIVACY OF DEFAULT APPS IN APPLE SMOBILE ECOSYSTEM Amel Bourdoucen.pdf

共23页,预览5页

还剩页未读，继续阅读

声明：本站为文档C2C交易模式，即用户上传的文档直接被用户下载，本站只是中间服务平台，本站所有文档下载所得的收益归上传人(含作者)所有。玖贝云文库仅提供信息存储空间，仅对用户上传内容的表现方式做保护处理，对上载内容本身不做任何修改或编辑。若文档所含内容侵犯了您的版权或隐私，请立即通知玖贝云文库，我们立即给予删除！

SYSTEMATIC EVALUATION AND USERSTUDY OF PRIVACY OF DEFAULT APPS IN APPLE SMOBILE ECOSYSTEM Amel Bourdoucen

相关推荐

开通VIP享超值会员特权

作者详情

相关内容

热门标签

举报选择: