A User-Centric Context-Aware Permission Governance Framework for Privacy Control in Default Mobile Applications

A User-Centric Context-Aware Permission Governance Framework for Privacy Control in Default Mobile Applications Asmau Yetunde Adeniran∗

asmauadeniran2004@gmail.com Department of Cybersecurity Air Force Institute of Technology Kaduna, Nigeria

arXiv:2608.27914v1 [cs.CR] 28 Aug 2026

Samaila Musa Abdullahi

Department of Cybersecurity Air Force Institute of Technology Kaduna, Nigeria

Adeniran Kolade Ademuwagun

Fatimah Adamu-Fika

Department of Cybersecurity Air Force Institute of Technology Kaduna, Nigeria

Department of Cybersecurity Air Force Institute of Technology Kaduna, Nigeria

Freeman Bitrus

Fortune Daberechi Ifeanyi

Department of Cybersecurity Air Force Institute of Technology Kaduna, Nigeria

Department of Cybersecurity Air Force Institute of Technology Kaduna, Nigeria

Abstract

CCS Concepts

Mobile operating systems offer runtime permission systems that are meant to improve the user control over access to sensitive data. Nevertheless, default or pre-built mobile applications are commonly much more deeply embedded into the system and can run with high privilege and cannot be scrutinized by the user. Current models of permission normally permit access in a persistent or a temporary state as long as an application session is alive, but do not differentiate between different features of the application. Consequently, users can give authorizations with little understanding of at what time and why certain information is being accessed. This paper presents a context-sensitive and user-focused permission governance system that would enhance privacy management in default mobile applications. The framework proposes a featurebased authorization option known as allow when needed which only allows access to a data to the functionality that needs it and not to the entire application session. A weighted scoring system was created to determine the privacy implication of the choices made by users, which was founded on the sensitivity of permission and the type of authorization. To simulate 30 realistic permission request situations in six default application types that are commonly used, a web-based simulation platform used to enable controlled, early-stage evaluation prior to native implementation platform was put in place. The assessment methodology adopts an exploratory approach using a cross-sectional survey comprising of 104 respondents to test the privacy awareness and behavior concerning the default applications and to supplement with formative usability testing consisting of eight respondents to test the interaction with the prototype system. According to the results of the surveys, not all users always elaborate the default app permission and would prefer that it be explained in a more contextual way before they could adopt it. These findings provide preliminary evidence that context-aware permission governance can improve user understanding and decision clarity during permission interactions. Given the limited availability of user-centric evaluations specifically targeting default mobile applications, this study adopts an exploratory simulation-based approach as an appropriate first step toward examining feature-level authorization and privacy feedback mechanisms before native mobile deployment.

• Security and privacy → Usability in security and privacy.

Keywords Mobile phone privacy; cybersecurity; Default mobile applications; user privacy awareness; permission abuse; Context-aware authorization; permission management

1

Introduction

Mobile phones are now vital computing gadgets of communication, navigation, access to information, health care, and daily digital transactions. Most of this is through default mobile applications included with the operating system, including browsers, messaging apps, maps, cameras, phone services, and health utilities. The default applications often seek permission to access sensitive resources, such as location, microphone, camera, contacts, storage and notifications in order to deliver these services. Although this kind of access facilitates convenience and usability, it also poses privacy threat where the access may not be easily visible, restricted, and interpreted by the user. Felt et al. examined user attention to and comprehension of mobile platforms and reported that users often struggle to accurately evaluate the privacy implications associated with application permissions [6]. Mobile operating systems attempt to mitigate these risks through permission-based access control mechanisms, which allow users to grant or deny applications access to protected system resources. Over time, these systems have evolved from install-time permission models toward runtime permission systems that allow users to make authorization decisions during application use. Wijesekera et al. demonstrated that permission decisions depend on the context in which access occurs, emphasizing the importance of contextual and understandable permission prompts [21]. Despite these improvements, existing permission systems largely operate at the application or session level rather than the feature level. Once a permission is granted, access may persist for the duration of the application session even when the specific feature that required the permission is no longer active. Mendes et al. conducted a field study examining user expectations regarding mobile

∗ Corresponding author.

1

3. A simulation-based implementation and exploratory empirical evaluation examining how feature-level authorization structures influence user awareness and decision-making behavior during permission interactions.

application privacy and found that many permission accesses occur in ways that users do not anticipate, resulting in discrepancies between user expectations and system behavior [12]. Privacy concerns are particularly significant for default or preinstalled mobile applications, which are deeply integrated into the operating system and often receive higher levels of implicit user trust compared to third-party applications. Ozbay and Bicakci proposed a security and privacy risk scoring model for pre-installed Android applications and observed that such applications may operate with privileged access while receiving less scrutiny from users than applications downloaded from official app stores [15]. Similarly, Sutter et al. analyzed firmware images of Android devices and demonstrated that pre-installed applications may include embedded third-party components and operate under advantageous system conditions, which can make their behavior difficult for users to monitor or evaluate [18]. This paper proposes a context-aware, user-centric permission governance framework designed specifically for default mobile applications, where permission behavior is often less visible to users despite deeper integration with operating system services. Unlike conventional runtime permission approaches that maintain access across entire application sessions, the proposed framework introduces a feature-level authorization option called “Allow When Needed,” which restricts permission access to the specific functionality that requires it rather than the duration of application execution. In addition, the framework incorporates a weighted privacy scoring model that provides interpretable feedback about cumulative exposure resulting from repeated permission decisions across interaction scenarios. This scoring mechanism is intended to support user awareness during authorization rather than enforce systemlevel policy restrictions. Because modifying native mobile operating systems was outside the scope of this study, the proposed approach was implemented and evaluated via a web-based simulation platform to enable controlled early-stage assessment before native mobile deployment thus emulating realistic permission request scenarios in default applications The evaluation consists of a cross-sectional survey involving 104 participants to examine user awareness and behavioral patterns regarding default application permissions, combined with formative usability testing with eight participants to explore how users interpret and interact with feature-level authorization and privacy feedback mechanisms within controlled permission-request scenarios. The contributions of this work are threefold and focus on exploratory evaluation of user-centered permission governance mechanisms for default mobile applications:

The remainder of the paper is organized as follows. Section 2 reviews related work on mobile permissions and context-aware authorization. Section 3 presents the proposed framework and simulation design. Section 4 reports the evaluation methodology and results. Section 5 discusses implications and limitations, and Section 6 concludes the paper.

2 Literature Review 2.1 Mobile Permission Models And Runtime Controls Access to sensitive resources like location, camera, microphone, and storage is controlled with the help of runtime permission frameworks, which are found in mobile operating systems. According to the Android Developers documentation, the latest Android operating systems have such options as one-time access and a possibility to allow only when using the app, which is directed to minimize the ongoing data exposure [2]. These mechanisms represent a shift from earlier install-time permission models toward more interactive user control. Although these have been improved, user control remains mostly at the application or session level. Mendes et al. performed a field study on user expectations during permission requests in actual field settings and discovered that a substantial part of the permission access was unexpected by the user, which revealed that there is a discrepancy between the intent of the user and the behavior of the system [12]. This issue of persistence comes into play particularly when permissions that have been granted to one feature are still persisted throughout the entire application session.

2.2

User Awareness and Permission Decision Behavior

Permission governance is highly dependent on user awareness. A study by Prange et al. investigated the awareness of the granted permission among Android users and found that a significant portion of respondents failed to correctly identify which permissions were granted to their installed applications [16]. This implies that permission states tend to vanish when one simply uses the phone in their normal daily use. The decision-making is also influenced by the quality of permission explanations. Elbitar et al. proved that the specifics of permission rationale words and the way they are presented play a crucial role in shaping the decisions of the users and their belief in the decisions made [5]. Their results support the assertion that permission systems are not technical controls but also communication, in which clarity and context determine user behavior. Wu et al. also delved into the issue of how user discomfort concerning permissions is dependent on physical contexts and activities [23]. Their findings show that the perception of privacy risk by users is situational, and it is possible to argue that the static or session-based models of permission are not sufficient to describe the dynamic expectations of users.

1. A feature-based permission governance model tailored for default mobile applications that links authorization decisions to individual functional interactions rather than application sessions. 2. A cumulative privacy scoring mechanism designed to provide interpretable feedback about relative exposure associated with repeated permission decisions across application scenarios. 2

2.3

populations of applications and may cause unwanted behaviors when permissions are revoked or permission specifications change across Android versions. This argument motivates the practical issue that though control mechanisms may be enhanced in platforms, actual usage may not conform to the intention of the user because of the actualities of implementation across platforms and OS versions [20]. On the ecosystem level, Alkinoon et al. examined the usage trends of permission over the years and across the app types and found that the patterns of permission requests change over time and that co-occurrence of permission sets can uncover significant privacy implications. This underlines the necessity of governance methods that can be interpreted by consumers alongside that which withstands the variation of applications through updates [1]. On the developer side developer and end user views of permissions were compared and, according to Tahaei et al. developers can request more substantial permissions because they have not understood permission scopes, or because the third party libraries require it, whereas users usually interpret the prompt to mean they are needed. This failure of connection assists us in understanding why users still encounter inappropriate demands even in contemporary runtime permission systems, and it inspires strategies of linking authorization more clearly to feature particular need instead of the broad access of the app as a whole [19].

Fine-Grained and Context-Aware Enforcement Approaches

Having identified the shortcomings of coarse-grained permission models, various researchers have considered context-sensitive and fine-grained enforcement system. Guerra et al. presented RPCDroid, a runtime analysis system, that correlates the identifiable dangerous use of permissions with particular UI events, making it possible to identify various access contexts within the same application [7]. Their contribution draws our attention to the fact that what is suitable to be permitted in one feature can be overly permitted in another. Likewise, Malviya et al. presented DROIDGEM which categorizes the permission usage on the per-functional level of a particular app as a response to user or system events [11]. This strategy facilitates the interpretation of permission access features on a per-feature basis and not on an application as a unit. Most recently, Milanese et al. tested a permission interpretation model based on the use of an LLM, which interprets UI context to aid in providing a user with a decision-making process [13]. Their article is explicit that existing mobile platforms are based on permissions on an application level instead of limiting them to particular features, allowing them to use more granular governance models.

2.4

Privacy Concerns In Default And Pre-Installed Applications

2.6

There are unique aspects of privacy, with default and pre-installed applications, due to their thorough access to the operating system and their special privileges during deployment. By analyzing firmware images of large-scale firms using a non-executable, largescale, and unbiased method, Sutter et al. revealed that pre-installed Android apps may contain hefty third-party abbreviates and function under advantageous circumstances [18]. These attributes may make user monitoring complex. Ozbay and Bicakci suggested a scoring system to evaluate the privacy and security risk of pre-installed Android applications, and stated that pre-installed applications tend to be less scrutinized than third-party applications downloaded in official stores [15]. Innovative analyses of mobile devices within the emerging markets have also documented privacy exposure associated to bundle software arrangement and privilege settings . Even though the vendors of the platform still enhance the privileged permission restrictions, such as allowlisting signature permissions of recent versions of Android, the changes happen mostly at the system level. They fail to directly discuss how user experience and interpret permission decisions in their daily interaction with default applications.

2.5

Prompting in Context, Rational Reasoning and Permission Fatigue

In addition to awareness, the time and manner of permitting influences quality of decisions. Harbach et al. conducted research on the large scale quieting of permission prompts in an on device environment and demonstrated that the reduction of unnecessary interruptions can enhance user experience without removing user control. Even though their context is browser permissions, the mechanism underlying them is applicable to the context of mobile permission governance as it demonstrates the conflict between overprompting and significant consent [8]. Elbitar et al. conducted a detailed study of permission rationales and assessed the effectiveness of the phrasing of the rationale in determining the user choice and perceived control. Their results confirm the practical design requirement that permission prompts should convey the immediate purpose and context, particularly when users cannot determine an individual feature to a permission prompt. This is similar to context anchored authorization, in which explanations are provided in addition to what the user is executing, rather than what the application can execute [5].

2.7

Design Implication of Permission Misuse, Ecosystem Evolution, and Design Implications

Fine Grained Projections between UI Context and Permissions

An accumulating literature tries to relate permissions to the particular interaction situation which provokes access to resources. Guerra et al. presented the dynamic analysis tool RPCDroid that associates UI-based events with permissions to access resources sensitive and proved that applications can use the same sensitive resource with varying user interactions. This helps in the major assumption that a single session level grant can be too broad since

Recent work also demonstrates that the permission problem is not a user interface problem, but an ecosystem evolution problem as well. Wang et al. studied bugs related to Android Runtime Permission and discovered that these bugs are typical among large 3

a user would mean granting one feature and not others that are using the same session window [7]. Milanese et al. examined the fact that large language models can determine the suitability of permission requests based on UI screen context, as they essentially view the screen as evidence of an active feature. Their findings substantiate the practicability of feature conscious permission judgement, and offer outward validation of forms of governance scheme that constrain authorization to the active functionality [13]. Independently, Hu et al. suggested an LLM driven API permission mapping discovery in the Android framework, which is also motivated by the observation that incomplete or obsolete permission mappings may result in wrong permission declaring statements. Though this work is aimed at finding the right answer, it is applicable to user governance since erroneous permission declaration renders user facing decisions meaningless, even with a well-designed UI [9].

2.8

often occurred in bursts and primarily affected rarely used apps or permissions that were not essential to core functionality [16]. The practice further supports the thesis that it might be necessary to consider other mechanisms of retrospective review.

2.10

A number of research papers have been done on adaptive systems to help users decide on permission. The contextual recommendation mechanisms presented by Olejnik et al. make use of the previous user decisions to infer authorization preferences in the future [14]. Their findings show that adaptive systems are able to reduce cognitive load without destroying user agency. On the same note, Wijesekera et al. compared machine learning models that make permission appropriateness decisions based on contextual metadata and history of user interactions [22]. These systems are also architecturally distinct from feature-level authorization, but have the same purpose: eliminating the distance between what is requested and what the user expects. Recently, Liu et al. found that 78.7% of the recommendations made by their personalized privacy assistant were adopted and that participants perceived the assistant and its recommendations as useful and usable [10]. This promotes the need for explainability coupled with enforcement.

Software and Firmware Level Transparency Default

Transparency, where default and pre-installed applications are concerned, needs more than app store level analysis, it needs firmware level visibility. Sutter et al. introduced FirmwareDroid to perform large scale extraction and static analysis of pre-installed Android software to restore the transparency of what is shipped on devices and the behavior of embedded software components. This clarifies the point that default applications should be given special treatment in terms of governing since their existence and actions may be more difficult to monitor or eliminate using normal tools [18]. Diallo et al. have explored what they call pre-installed android apps on low cost devices and have discussed how very cheaply priced phones can become a distribution channel of privacy violating functionality particularly in areas where such phones are widespread. This further inspires permission governance designs which do not presuppose app store inspection and which work well even when the app is a default unit of the device environment [4].

2.9

Adaptive Privacy Controls and Decision Support Systems

2.11

Governance in Transparency Mechanisms and Platforms

The tools of platform level transparency have also developed. Prange et al. found that most participants had previously used Android’s Permission Manager, though usually less than once a month, while most Android 12 participants had not used or were unaware of the Privacy Dashboard [16]. This is in concurrence with the evidence in the past where awareness may not necessarily translate into long term governance behavior. Apple App Privacy labels have also been evaluated in the same way. According to Xiao et al., there were gaps between the disclosed data practices and the network behavior observed in some types of applications [24]. Based on these results we can conclude that the declarative transparency mechanisms are not capable of ensuring meaningful control.

Permission Exposure and Behavioral Drift Empirical Measurement

The complicated nature of permission exposure in modern mobile ecosystems is still evidenced by large-scale empirical studies. Ren et al. investigated network-based personal information leaking in mobile traffic and demonstrated that the leakage may happen even when legitimate applications transmit personal information, but in the form not intuitive to users [17]. Although the concepts are technically applied to default applications, more generally this work indicates the issue of matching the perceived permission grants to the reality of downstream data flows. Calciati and Gorla examined updates to Android applications using a longitudinal study and stated that permission requests tend to widen with time, especially after major feature releases [3]. This evolutionary trend makes user management more complicated since an authorization that had been issued with scaled functionality might subsequently be extended to new behaviors. Prange et al. examined users’ awareness and control of Android privacy permissions in a field study and found that revocations

2.12

Research Gap and Positioning Of This Study

Recent research efforts have also made significant progress in enhancing mobile permission systems with runtime controls, contextual explanations, and fine-grained analysis of permission usage. Past work on mobile permission systems has also covered aspects of context-aware authorization mechanisms, feature-level permission mapping, adaptive decision support systems, and other aspects that assist users during permission interaction. Overall, all past work on mobile permission systems indicates that permission interaction is closely related to user expectations in terms of context relevance. However, there are certain limitations associated with past work on mobile permission systems in terms of user-centric permission governance. First, much of past work on mobile permission systems 4

is more focused on system-level or developer-level permission analysis, permission detection, or permission decision support. Therefore, there is less emphasis on user-centric permission governance. Second, although fine-grained permission analysis is part of past work on mobile permission systems, there is less emphasis on integrating permission governance with user-centric features in terms of permission authorization.Third, default or pre-installed mobile applications, despite their higher privileges and popularity, are an area that is still unexplored with regard to user-facing permission governance structures. The current study seeks to fill in the identified research gaps by developing a user-centric permission governance model that incorporates three components within a unified interaction model, namely, feature-level permission governance through the “Allow When Needed” option, contextual permission explanations in association with functional interactions, and cumulative privacy scores with the intention of offering interpretable feedback to the user on their interactions with the permission governance framework. Rather than developing a new system-level enforcement mechanism, the novelty of the current study is in developing a unified user-facing interaction model that is specific to default mobile applications. Considering the challenges in directly implementing such permission governance structures within mobile operating systems, the current study seeks to offer an alternative simulation-based approach to understanding user interactions with feature-level permission governance structures, with the intention of offering an understanding of how such structures can influence user awareness in association with their interactions with the permission governance framework.

Figure 1: Overview of the Research Design and Data Collection Workflow

3 Methods 3.1 Research Approach

Figure 2: System Architecture and Component Integration

The paper utilized a system design and user evaluation technique to analyze privacy control measures in default mobile applications. The study entailed creation of a web-based simulation system deploying a situation-aware permission governance model and exploratory empirical evaluation with survey data collection and formative usability testing. The aim of the methodological design was to determine the effect of the feature-level authorization and privacy scoring feedback on the user awareness and decision behavior in the controlled permission request scenarios. The overall research process involved parallel data collection and system development phases. Survey data were collected to examine user awareness and behavioral patterns, while the simulation platform was developed to evaluate feature-level authorization within controlled scenarios. The workflow of the study is illustrated in Fig. 1. As shown in Fig. 1, survey analysis informed interpretation of user behavior patterns, while mock scenario datasets were used to construct realistic permission interactions within the simulation environment.

3.2

System Design and Implementation

The simulation platform is a web-based application that was created to represent permission interactions that are a common occurrence in default mobile applications. The system was developed in such a manner that it imitates realistic permission request prompts and can compare authorization options in a structured manner. The system consists of a scenario module responsible for presenting predefined application tasks, a permission decision interface that provides four authorization options, a privacy scoring module that updates cumulative exposure metrics, and a data recording component that logs user interaction for evaluation purposes. The internal architecture of the web-based simulation platform is composed of a backend logic layer, frontend interface layer, encryption handling module, and integration mechanism. The structural relationship between components is illustrated in Fig. 2. The backend was implemented using Python and Flask, the frontend utilized HTML, CSS, and JavaScript, and encryption operations were handled using AES via the Python cryptography library. These 5

The cumulative privacy score was successively updated upon each scenario interaction. The scoring system was intended to give relative feedback in the simulation environment, and to represent relative exposure to privacy in various authorization options. The model is not an absolute privacy risk but works as an assessment measure to aid user awareness in interaction.

3.5

Scenario Construction

To test the framework, structured permission scenarios were created on the basis of frequent interactions with default mobile applications. Six default applications were selected for scenario construction, including Browser, Messages, Maps, Camera, Health, and Phone. Each application was characterized by 5 task-based scenarios, which made a total of 30 simulated permission interactions. Each scenario consisted of a functional task description, the associated permission request, four authorization options, and a contextual explanation describing the rationale for access. The scenario set was to be created to have selected sensitive permissions such as location access, microphone, camera, contact access, storage access, and notification permissions. The constructed scenarios focused on foreground permission requests triggered by explicit user interactions with application features, while background or passive permission access behaviors were not included within the scope of the simulator-based evaluation. Figure 3: Permission Authorization Options in the Simulation Interface

3.6

components were integrated to ensure secure and structured handling of scenario execution and user interaction logging.

3.3

Permission Options and Feature-Level Authorization

The simulation platform offers four permission choices to the authorization of every permission request: “Don’t Allow,” “Allow While Using App,” “Allow When Needed,” and “Always Allow.” The “Allow When Needed” option represents the feature-level authorization mechanism proposed in this study. This option is a conceptual limit in contrast to the session-based authorization, that preserves permission access during an application session. Figure 3 illustrates the four authorization options presented to users during each simulated permission request.

3.4

3.7

Participant Demographics

The demographic characteristics of survey participants were analyzed to support interpretation of permission awareness and behavioral response patterns. A total of 104 valid responses were obtained from users representing multiple age groups and varying levels of familiarity with smartphone privacy settings. The majority of respondents (79 participants) were between 18 and 24 years of age, followed by participants aged 25–30 (13 participants), under 18 (5 participants), 35–44 (4 participants), and 45–54 (3 participants). This distribution reflects a participant population actively engaged in smartphone usage and representative of common mobile application users. In terms of familiarity with privacy settings, 60 participants reported being very comfortable managing permission controls, 36 participants indicated basic familiarity, and 8 participants reported limited familiarity. This variation supported evaluation across users with different levels of technical confidence.

Privacy Scoring Mechanism

A direct mathematical model was applied to measure the effects of permission decisions. A score change was calculated in each permission interaction according to the sensitivity weight of the permission requested and the selected authorization modifier. The score update mechanism is defined as: Score Change = Permission Weight × Action Modifier

Survey Design and Data Collection

To test the awareness of the user, trust perception and behavioral pattern about default mobile application permissions, a crosssectional survey was carried out. The survey was conducted online with 104 valid responses. The questionnaire contained demographic details, habits of device usage, habits of examining the approval, confidence in default applications, and the choices of descriptive permission explanations. They had closed-ended and open-ended questions. No personal identifiable information was gathered.

(1) 6

Table 1: Awareness of Default Application Permissions (n = 104) Response Yes No

Frequency (n)

Percentage (%)

75 29

72.1 27.9

Survey data were analyzed using descriptive statistical methods such as frequency distributions and percentage-based summaries to characterize user awareness, permission review behavior, trust perceptions, and preferences for contextual explanations. This approach was selected to provide an initial understanding of user attitudes and behavioral tendencies related to default mobile application permissions. In parallel, formative usability testing was conducted with eight participants to examine how users interpret and interact with the proposed authorization model and privacy scoring mechanism. The usability study is intended as a formative evaluation aimed at identifying interaction patterns, usability challenges, and user interpretation of feature-level authorization, rather than providing statistically generalizable results. The sessions focused on qualitative observation of user behavior, including comprehension of permission prompts, interpretation of the “Allow When Needed” option, and responses to cumulative privacy feedback.. The use of a web-based simulation platform enabled controlled presentation of permission scenarios across different application contexts, ensuring consistency in interaction conditions during evaluation. However, the findings should be interpreted as preliminary, as the simulated environment does not fully replicate real-world mobile usage conditions.

Figure 4: showing the age distribution of participants

Figure 5: showing the technical expertise of participants

4 Result 4.1 Overview of Respondents The following results present descriptive findings from the evaluation of 104 valid responses. Every respondent filled the main parts of the questionnaire related to awareness, review behavior, trust perception, and permission management preferences. There were no unfinished answers that were not used in statistical reporting.

Figure 6: showing the device ecosystem representation Regarding device ecosystem representation, 57 participants reported using Android devices, 52 reported using iOS devices, and 1 participants reported using both platforms, while 1 participant reported using a button phone. Inclusion of participants across major mobile operating systems strengthens the generalizability of the findings related to permission awareness and interaction behavior.

3.8

4.2

Awareness of Default Application Permissions

Table 1displays the distribution of answers in terms of awareness of permissions given to default or pre-installed mobile applications. Most of the respondents (72.1) stated that they knew the permissions they gave to default applications on their devices. Nevertheless, awareness does not necessarily constitute active privacy management behavior as discussed in the analysis below.

Evaluation Procedure

The evaluation was designed to assess user awareness, behavioral patterns, and interaction with feature-level permission governance within a controlled environment. Given the absence of prior usercentric evaluation frameworks specifically addressing permission governance in default mobile applications, an exploratory design was adopted as an appropriate first step. The analysis therefore focuses on identifying observable trends and usability insights rather than establishing actual causal relationships or statistically generalizable conclusions.

4.3

Permission Review and Revisit Behavior

Table 2 presents the frequency of default application permissions review or modifications by the respondents. Awareness levels were rather high, but there was a lack of consistent review behavior. Just 11.5% of the respondents said they would review permissions frequently or regularly. In contrast, 58.7% 7

Table 2: Frequency of Permission Review (n = 104) Review Frequency

Frequency (n)

Percentage (%)

4 8 31 47 14

3.8 7.7 29.8 45.2 13.5

Always (after update) Often (weekly) Sometimes (monthly) Rarely (≈ every 6 months) Never

Figure 8: Trust Levels in Default Applications trust and reported discomfort point to variability in user perception in the process of the permission interaction. Figure 5 provides a graphical representation of reported trust levels in default mobile applications.

4.5 Figure 7: Frequency of Permission Review Among Respondents

Table 5 summarizes participants’ preferences that pertain to contextual explanations and reminder mechanisms for permission management. Percentage was calculated among those respondents who answered the item. An overwhelming majority of respondents indicated that contextual explanations would improve their comfort when granting permissions. Similarly, most respondents indicated a desire for reminders or alerts for managing default application permissions effectively. These results highlight that the users value proactive communication about permissions and prefer mechanisms that provide clarity before they make any decisions.

Table 3: Revisit of Permission Settings After Updates (n = 104) Revisit Frequency Always Sometimes Rarely Never

Frequency (n)

Percentage (%)

3 24 40 37

2.9 23.1 38.5 35.6

4.6

reported either rarely or never checking default application permissions. To visually illustrate the distribution of permission review behavior, Fig. 4 presents the percentage breakdown of respondents’ reported review frequency. The visualization reinforces the concentration of responses in the “Rarely” and “Never” categories, supporting the observation that sustained permission monitoring is limited among users. Table 3 shows revisit behavior before and after application or system updates. Over two-thirds of the participants (74.1%) said they would hardly or never revisit permission settings once updated. This trend indicates the lack of behavioral consistency between first-time awareness and continuous monitoring of permission settings.

4.4

Preference for Contextual Explanations and Alerts

Self-Reported Confidence in Permission Management

The responses indicate a variation in user confidence when managing permissions for default mobile applications. Although many participants did report general awareness of permission settings, this did not consistently translate into confidence during real-time decision-making. Observations from the usability sessions suggest that uncertainty often arises at the moment of permission request, particularly when the purpose and scope of access are not clearly tied to a specific feature. The participants demonstrated higher confidence when the permission requests were directly associated with an active task and when feedback mechanisms provided an immediate interpretability of their choices. These findings suggest that confidence in permission management is influenced not only by prior knowledge of settings but also by the clarity and timing of said permission interactions, supporting the role of feature-level authorization in reducing uncertainty when decisions are being made.

Trust in Default Applications and Permission Discomfort

Table 4 is a summary of trust perception and the discomfort experienced in the pre-request period in the permission requests. The majority respondents have moderate trust, with 63.5% having moderate trust in default applications. Simultaneously, 61.5% indicated that they have felt uneasy about a permission that had been asked by a default application. The combination of moderate

4.7

Consolidated Observations

Taken together, the results reveal an important gap between user awareness and sustained privacy management behavior. While a 8

Table 4: Trust and Discomfort Regarding Default Applications (n = 104) Variable

Response Category

Trust Level

Completely A little Not at all Yes No

Experienced discomfort about a permission

Frequency (n)

Percentage (%)

19 66 19 64 36

18.3 63.5 18.3 61.5 34.6

Table 5: Preference for Contextual Explanations and Permission Alerts Variable

Response

Frequency (n)

Percentage (%)

Contextual explanation would enhance the comfort Reminders or alerts would be desirable

Yes Yes Not sure No

60 81 18 1

93.8 77.9 17.3 1.0

large proportion of respondents reported being aware of permissions granted to default applications, most indicated that they rarely revisited permission settings after updates or during routine device usage. This pattern suggests that awareness alone does not translate into consistent privacy-protective behavior. At the same time, participants expressed moderate trust in default applications alongside noticeable levels of discomfort when permission requests were presented without sufficient explanation. The strong preference for contextual explanations and reminder mechanisms further indicates that users are more comfortable interacting with permission systems that provide feature-level transparency rather than persistent session-level authorization. These observations collectively support the motivation for the proposed “Allow When Needed” framework, which introduces contextual permission triggers and explanation-aware authorization prompts to improve user engagement with privacy decisions at the feature level.

5

authorization decisions in practice. The feature-level authorization model proposed in this study intended to address this mismatch by linking permission activation to individual functional interactions rather than maintaining access across entire application sessions. Although the evaluation was conducted within a simulation environment rather than a native operating system implementation, the findings provide preliminary evidence that context-sensitive authorization logic may align more closely with user expectations and could improve perceived control over privacy-related decisions. This research paper has several limitations that should be considered when interpreting the findings. First, the results rely partially on self-reported survey responses, which may not always reflect actual user behavior during real-world permission decision-making. In addition, although the sample size was appropriate for descriptive behavioral analysis and exploratory usability evaluation, broader validation with larger and more diverse participant populations would further strengthen the generalizability of the findings. Additionally, the complete reliance on descriptive statistical analysis limits the ability to draw causal inferences or statistically and reliably validate observed differences in user behavior. A further limitation relates to the use of a web-based simulation environment rather than deployment within a native mobile operating system. Because participants interacted with permission scenarios in a controlled experimental setting, their decision-making behavior may have been influenced by awareness of being observed as part of a study. This observer effect may lead users to act more cautiously than they would during routine smartphone use, particularly when responding to permission requests involving sensitive data access. Future work will involve implementing the proposed framework within a native mobile environment and conducting actual controlled experiments using inferential statistical methods to evaluate its impact on user decision-making and privacy outcomes under real-world conditions. In addition, the simulated scenarios focused primarily on foreground permission requests triggered by explicit user interactions with application features. However, many default mobile applications perform background data access operations without direct

Discussion

The results of this research indicate that there is a consistent pattern in user behavior regarding permission management in default mobile applications. Most respondents reported that they were aware of the permissions granted to pre-installed applications; however, sustained review and active management of these permissions remained limited. This suggests that awareness alone does not necessarily translate into continuous privacy governance behavior. At the same time, a significant number of respondents reported discomfort with particular permission requests despite maintaining moderate confidence in default applications overall. This indicates that user concern is not primarily driven by general distrust, but rather by uncertainty regarding the scope, duration, and necessity of permission access in specific situations. The preference expressed by participants for contextual explanations and reminder mechanisms further emphasizes the importance of delivering permission information at the moment of decisionmaking. These findings highlight a structural mismatch between existing permission models and the way users interpret and manage 9

user awareness, including passive location tracking, synchronization processes, and system-level telemetry collection. These forms of invisible permission usage represent an important dimension of privacy exposure that was not captured within the scope of the present simulator-based evaluation. Future extensions of the proposed framework will incorporate monitoring mechanisms for background permission activity in order to support more comprehensive protection against passive data access by default applications.

6

[7] Michele Guerra, Roberto Milanese, Rocco Oliveto, and Fausto Fasano. 2023. RPCDroid: Runtime Identification of Permission Usage Contexts in Android Applications. In Proceedings of the 9th International Conference on Information Systems Security and Privacy (ICISSP 2023). SciTePress, 714–721. doi:10.5220/ 0011797200003405 [8] Marian Harbach, Igor Bilogrevic, Enrico Bacis, Serena Chen, Ravjit Uppal, Andy Paicu, Elias Klim, Meggyn Watkins, and Balazs Engedy. 2024. Don’t Interrupt Me— A Large-Scale Study of On-Device Permission Prompt Quieting in Chrome. In 31st Annual Network and Distributed System Security Symposium (NDSS 2024). Internet Society. https://www.ndss-symposium.org/ndss-paper/dont-interrupt-me-alarge-scale-study-of-on-device-permission-prompt-quieting-in-chrome/ [9] Han Hu, Wei Minn, Yonghui Liu, Jiakun Liu, Ferdian Thung, Terry Yue Zhuo, Lwin Khin Shar, Debin Gao, and David Lo. 2025. Bamboo: LLM-Driven Discovery of API-Permission Mappings in the Android Framework. arXiv:2510.04078 [cs.SE] doi:10.48550/arXiv.2510.04078 [10] Bin Liu, Mads Schaarup Andersen, Florian Schaub, Hazim Almuhimedi, Shikun (Aerin) Zhang, Norman Sadeh, Yuvraj Agarwal, and Alessandro Acquisti. 2016. Follow My Recommendations: A Personalized Privacy Assistant for Mobile App Permissions. In Twelfth Symposium on Usable Privacy and Security (SOUPS 2016). USENIX Association, Denver, CO, 27–41. https: //www.usenix.org/conference/soups2016/technical-sessions/presentation/liu [11] Vikas K. Malviya, Yan Naing Tun, Chee Wei Leow, Ailys Tee Xynyn, Lwin Khin Shar, and Lingxiao Jiang. 2023. Fine-Grained In-Context Permission Classification for Android Apps Using Control-Flow Graph Embedding. In 2023 38th IEEE/ACM International Conference on Automated Software Engineering (ASE). IEEE, 1225– 1237. doi:10.1109/ASE56229.2023.00056 [12] Ricardo Mendes, André Brandão, João P. Vilela, and Alastair R. Beresford. 2022. Effect of User Expectation on Mobile App Privacy: A Field Study. In 2022 IEEE International Conference on Pervasive Computing and Communications (PerCom). IEEE, 207–214. doi:10.1109/PerCom53586.2022.9762379 [13] Roberto Milanese, Michele Guerra, Michele Daniele, Giovanni Fabbrocino, and Fausto Fasano. 2025. Assessing the Effectiveness of an LLM-Based Permission Model for Android. In Proceedings of the 11th International Conference on Information Systems Security and Privacy (ICISSP 2025), Vol. 2. SciTePress, 36–47. doi:10.5220/0013128100003899 [14] Katarzyna Olejnik, Italo Dacosta, Joana Soares Machado, Kévin Huguenin, Mohammad Emtiyaz Khan, and Jean-Pierre Hubaux. 2017. SmarPer: Context-Aware and Automatic Runtime-Permissions for Mobile Devices. In 2017 IEEE Symposium on Security and Privacy (SP). IEEE, 1058–1076. doi:10.1109/SP.2017.25 [15] Abdullah Özbay and Kemal Bıçakcı. 2024. Should Users Trust Their Android Devices? A Scoring System for Assessing Security and Privacy Risks of PreInstalled Applications. ITU Journal of Wireless Communications and Cybersecurity 1, 1 (2024), 9–28. https://dergipark.org.tr/en/pub/itujwcc/article/1463270 [16] Sarah Prange, Pascal Knierim, Gabriel Knoll, Felix Dietz, Alexander De Luca, and Florian Alt. 2024. “I do (not) need that Feature!”—Understanding Users’ Awareness and Control of Privacy Permissions on Android Smartphones. In Twentieth Symposium on Usable Privacy and Security (SOUPS 2024). USENIX Association, Philadelphia, PA, 453–472. https://www.usenix.org/conference/ soups2024/presentation/prange [17] Jingjing Ren, Ashwin Rao, Martina Lindorfer, Arnaud Legout, and David Choffnes. 2016. ReCon: Revealing and Controlling PII Leaks in Mobile Network Traffic. In Proceedings of the 14th Annual International Conference on Mobile Systems, Applications, and Services (MobiSys ’16). Association for Computing Machinery, New York, NY, USA, 361–374. doi:10.1145/2906388.2906392 [18] Thomas Sutter and Bernhard Tellenbach. 2023. FirmwareDroid: Towards Automated Static Analysis of Pre-Installed Android Apps. In 2023 IEEE/ACM 10th International Conference on Mobile Software Engineering and Systems (MOBILESoft). IEEE, 12–22. doi:10.1109/MOBILESoft59058.2023.00009 [19] Mohammad Tahaei, Ruba Abu-Salma, and Awais Rashid. 2023. Stuck in the Permissions With You: Developer & End-User Perspectives on App Permissions & Their Privacy Ramifications. In Proceedings of the 2023 CHI Conference on Human Factors in Computing Systems (CHI ’23). Association for Computing Machinery, New York, NY, USA, Article 168. doi:10.1145/3544548.3581060 [20] Sinan Wang, Yibo Wang, Xian Zhan, Ying Wang, Yepang Liu, Xiapu Luo, and Shing-Chi Cheung. 2022. APER: Evolution-Aware Runtime Permission Misuse Detection for Android Apps. In Proceedings of the 44th International Conference on Software Engineering (ICSE ’22). Association for Computing Machinery, New York, NY, USA, 125–137. doi:10.1145/3510003.3510074 [21] Primal Wijesekera, Arjun Baokar, Ashkan Hosseini, Serge Egelman, David Wagner, and Konstantin Beznosov. 2015. Android Permissions Remystified: A Field Study on Contextual Integrity. In 24th USENIX Security Symposium (USENIX Security 15). USENIX Association, Washington, D.C., 499–514. https://www.usenix. org/conference/usenixsecurity15/technical-sessions/presentation/wijesekera [22] Primal Wijesekera, Arjun Baokar, Lynn Tsai, Joel Reardon, Serge Egelman, David Wagner, and Konstantin Beznosov. 2017. The Feasibility of Dynamically Granted Permissions: Aligning Mobile Privacy with User Preferences. In 2017 IEEE Symposium on Security and Privacy (SP). IEEE, 1077–1093. doi:10.1109/SP.2017.51

Conclusion

This paper has explored the ways in which feature-level authorization can be used to enhance permission governance of default mobile applications. An online simulation platform was created that incorporates an extra authorization policy, “Allow When Needed” as well as a privacy scoring system to give orderly feedback on permission choices. Optimal framework was tested on 30 simulated permission condition and research survey data of 104 mobile users. The findings indicate that the general awareness of default application permissions exists, but the behavior of active review is not prevalent, and still, there is a sense of discomfort when a particular permission is being requested. These results suggest that the current session-based permission models might not be adequate to meet the user expectations about the scope and duration of the access. The proposed model offers a more context-specific alternative to the conventional methods by limiting permission activation to feature interactions between individuals. Even though the evaluation was conducted within a controlled simulation environment, this study establishes an initial empirical and conceptual foundation for feature-level permission governance in default mobile applications. So, rather than presenting conclusive effectiveness, the findings highlight observable trends in user awareness, their interaction behavior, and preference for contextual authorization mechanisms. Future work will focus on implementing the framework within a native mobile environment and applying rigorous experimental and statistical methods to validate its impact under real-world conditions.

References [1] Ali Alkinoon, Trung Cuong Dang, Ahod Alghuried, Abdulaziz Alghamdi, Soohyeon Choi, Manar Mohaisen, An Wang, Saeed Salem, and David Mohaisen. 2025. A Comprehensive Analysis of Evolving Permission Usage in Android Apps: Trends, Threats, and Ecosystem Insights. Journal of Cybersecurity and Privacy 5, 3, Article 58 (2025). doi:10.3390/jcp5030058 [2] Android Developers. 2026. Request Runtime Permissions. Android Developers documentation. https://developer.android.com/training/permissions/requesting Accessed 2026-08-28. [3] Paolo Calciati and Alessandra Gorla. 2017. How Do Apps Evolve in Their Permission Requests? A Preliminary Study. In 2017 IEEE/ACM 14th International Conference on Mining Software Repositories (MSR). IEEE, 37–41. doi:10.1109/MSR. 2017.64 [4] Alioune Diallo, Anta Diop, Abdoul Kader Kabore, Jordan Samhi, Aleksandr Pilgun, Tegawendé F. Bissyande, and Jacque Klein. 2025. Security Evaluation of Android Apps in Budget African Mobile Devices. arXiv:2509.18800 [cs.CR] doi:10.48550/arXiv.2509.18800 [5] Yusra Elbitar, Alexander Hart, and Sven Bugiel. 2025. The Power of Words: A Comprehensive Analysis of Rationales and Their Effects on Users’ Permission Decisions. In 32nd Annual Network and Distributed System Security Symposium (NDSS 2025). Internet Society. doi:10.14722/ndss.2025.230544 [6] Adrienne Porter Felt, Elizabeth Ha, Serge Egelman, Ariel Haney, Erika Chin, and David Wagner. 2012. Android Permissions: User Attention, Comprehension, and Behavior. In Proceedings of the Eighth Symposium on Usable Privacy and Security (SOUPS ’12). Association for Computing Machinery, New York, NY, USA, Article 3, 14 pages. doi:10.1145/2335356.2335360 10

A

[23] Yuxi Wu, Jacob Logas, Devansh Ponda, Julia Haines, Jiaming Li, Jeffrey Nichols, W. Keith Edwards, and Sauvik Das. 2025. Modeling End-User Affective Discomfort With Mobile App Permissions Across Physical Contexts. In Symposium on Usable Security and Privacy (USEC 2025). Internet Society. https://www.ndsssymposium.org/ndss-paper/auto-draft-592/ [24] Yue Xiao, Zhengyi Li, Yue Qin, Xiaolong Bai, Jiale Guan, Xiaojing Liao, and Luyi Xing. 2023. Lalaine: Measuring and Characterizing Non-Compliance of Apple Privacy Labels. In 32nd USENIX Security Symposium (USENIX Security 23). USENIX Association, Anaheim, CA, 1091–1108. https://www.usenix.org/ conference/usenixsecurity23/presentation/xiao-yue

Open Science

Availability of Data and Materials: The survey dataset generated and analyzed during the current study is available from the corresponding author upon reasonable request

B

Ethical Considerations

Ethics Approval: This study involved voluntary survey participation. No personally identifiable information was collected. Conflicts of Interest: The authors declare no conflicts of interest to report regarding the present study.

11

Related documents

• A User-Centric Context-Aware Permission Governance Framework for Privacy Control in Default Mobile Applications#781192
• A User-Centric Context-Aware Permission Governance Framework for Privacy Control in Default Mobile Applications#779817
• The Not So Innocent Nudge: Why the Hooked Nudges for Smartphone Apps Undermine Autonomy#699202
• Factors Affecting Users' Disclosure Decisions in Android Runtime Permissions Model#329982
• Engineering Privacy in Smartphone Apps: A Technical Guideline Catalog for App Developers#964693