Achieving data privacy compliance presents a unique interdisciplinary challenge for experts from many backgrounds, particularly the technical and legal professions. As a potential solution for the legal mandate handed down by modern privacy regulations, Privacy-Enhancing Technologies (PETs) can serve as promising tools to help data processors demonstrate compliance. The implementation of PETs does not come immediately, however, and challenges in their adoption include their inherent technical complexity, as well as the lack of awareness and understanding of these technologies. In tackling these challenges, we investigate the educational needs of practitioners working in privacy compliance. Guided by Bloom's Revised Taxonomy, we begin the discussion on how the adoption of PETs can become more informed, with the goal of improving the efficiency and privacy consciousness of compliance programs. To accomplish this, we conduct 11 semi-structured interviews, analyze the results following Grounded Theory, and evaluate our findings in a survey with 24 respondents.
In a world where vast amounts of data are being created and processed on a continual basis, the
need for the responsible handling of such data has starkly risen. Along with increasing concerns
regarding the protection of individuals' privacy, the pressure placed on practitioners to comply
with relevant data privacy regulations such as the GDPR raises the stakes for data processors
[
Recently, the promise of Privacy-Enhancing Technologies (PETs) has saturated the academic
sphere, engaging researchers to develop innovative technologies for data privacy protection. In
essence, PETs encompass a range of technical approaches designed to protect the data of the
individual, when this data is utilized for some purpose. Such technologies, while falling under the
same class of Privacy-Enhancing Technologies, are highly diverse, particularly in their applicable
use cases. One unifying aspect, though, is their inherent complexity, which has kept their practical
adoption quite limited [
The road to widespread adoption of Privacy-Enhancing Technologies begins with the transition from PETs as a research topic to the dissemination of such knowledge to practitioners in the industry. However, essential questions then arise as to who constitutes the target audience, and what specific knowledge regarding PETs is required by practitioners. To identify the target
0000-0001-7485-2933 (A. Klymenko); 0000-0001-9230-5001 (S. Meisenbacher); 0000-0002-6667-5452 (F. Matthes) © 2023 Copyright for this paper by its authors.
Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
CEUR Workshop Proceedings (CEUR-WS.org)
audience, we look to the process of privacy compliance, which centers around the
implementation of appropriate technical measures for the safeguarding of personal data being
processed in a system. Gürses and Del Alamo [
In this work, we aim to investigate the educational needs of practitioners with respect to PETs,
with the goal of empowering them to be competent users of PETs, as "computer scientists and
particularly IT security experts with knowledge about privacy-enhancing technologies are
increasingly needed" [
A key step towards ensuring compliance with the data privacy regulations comes with the requirement to implement technical measures to protect the privacy of individuals. In this respect becomes important the concept of Privacy-Enhancing Technologies (PETs), a class of technologies that "protect privacy by eliminating or reducing personal data or by preventing unnecessary and/or undesired processing of personal data, all without losing the functionality of the information system" [9]. The recent guidance by the Information Commissioner’s Office (ICO) provides a detailed discussion on some of the prominent PETs, such as Differential Privacy, ZeroKnowledge Proofs, and Secure Multi-Party Computation, and outlines how they can help organizations to achieve data privacy compliance [10].
Although such advanced PETs present concrete solutions for personal data protection and
multiple real-world use case examples have been reported [11][12], they still remain
predominantly in the academic sphere and are not widely adopted in practice [
In this work, we consider the inherently interdisciplinary nature of privacy compliance and
investigate the educational needs of practitioners based on the different roles involved in the
process of privacy compliance, as proposed by Klymenko et al. [
To assess the educational needs of practitioners in learning about PETs, we designed the interview and survey studies to focus on extracting the learning goals of the questioned experts. In the interview, this was done in a semi-structured way, with two categories of questions: background questions, including the interviewee’s baseline knowledge of PETs, and questions aimed at identifying what kind of information about PETs is most relevant to the interviewee's role and responsibilities. A thematic content analysis according to Braun and Clarke [13] was conducted on the interview transcripts. The main goal of this analysis was to identify overarching themes expressed in the interviews, particularly relating to the learning needs and goals of practitioners with respect to PETs. Guided by following Grounded Theory (GT) Methodology [14], we analyzed interview transcripts concurrently to data collection and highlighted key themes, which were categorized into learning goals and educational needs. Axial coding was applied to identify relationships between these themes, supported by Bloom’s Revised Taxonomy.
Based on the resulting learning goals identified by our analysis and introduced in Table 3, the survey statements were designed to map learning goals to role-specific educational needs, where each statement corresponded to a cognitive process in Bloom’s Revised Taxonomy. The survey participants were then prompted to select the statement which best reflects their personal learning goals, allowing for the mapping of roles to levels in Bloom’s Revised Taxonomy.
Table 1 and Table 2 present relevant information on the interviewees and survey participants. Area identifies whether the survey respondents are working in a technical (T), legal (L), or GoBetween (G) role. Exp. represents years of relevant experience. To mitigate bias, no survey respondents also took part in the interview study.
To formulate and categorize the learning goals of practitioners regarding PETs, we employ
Bloom's Revised Taxonomy [
Factual knowledge includes terminology, characteristics, and features of PETs. The simplest learning goals are to list different PETs, as well as to know about the use cases of PETs, a topic most directly corresponding to Remember. Analyzing PETs on a factual level can be conceived as comparing different PETs and accordingly selecting technologies. It thus becomes clear that the tasks build up on each other, i.e., that Remember, Understand, and Apply are required to perform the subsequent Analyze tasks.
Conceptual knowledge is closely related to theoretical topics, such as introducing models, approaches, and interrelations of PETs. As opposed to factual knowledge, conceptual knowledge includes the principles behind the functionality of PETs. Based on the study results, statements are focused on system architecture, as interview participants reported a need to understand this topic better. The idea of integrating newly learned information into existing knowledge domains characterizes conceptual knowledge. However, it encapsulates the decision over which technology would be applicable; the implementation itself belongs strictly to the following category.
Here, the focus is placed on the implementation of PETs. Although the statements presented are expected to be universally applicable to all privacy roles, there is now a shift towards more technical content. Applying Procedural Knowledge marks the point where the learning content becomes rather technical, implying that a higher level of technical literacy is required. Furthermore, it shows how many learning goals can be identified before implementation. The next modification of the cognitive category is directed at the implementation action itself. The intent is not just to implement PETs in any fashion but to know parameters and quality measures, and thereby build an implementation strategy. Ultimately, the goal of procedural knowledge is not only to find the most suitable PET, but also to contribute to the development of new PETs.
Privacy-Enhancing Technologies are under constant pressure to evolve, as are any technologies employed to minimize risks or mitigate threats. The question of maturity is of great interest with Privacy-Enhancing Technologies. Achieving such knowledge requires a deep knowledge of the PETs in question, the environment in which PETs are implemented, and awareness of the limitations of the technologies. Therefore, learning goals in this knowledge category convey this critical approach, while also focusing on finding strategies to address these limitations. The highest learning goal would be to transfer knowledge to formerly unknown domains, identifying new purposes for PETs.
Table 3 presents the set of learning goal statements for PETs, which is based on Bloom’s Revised Taxonomy and supported by the interview findings. Using the guidelines provided by the original taxonomy and augmenting these with goals expressed by interviewees, we build the statements in Table 3 to align with the knowledge levels and cognitive processes of Bloom’s Revised Taxonomy. This mapping process is aided by Anderson and Krathwohl [15] and inspired by Servin et al. [16], the latter of which extends existing verb sets to include the technical domain.
Evaluate Create I want to verify I want to be statements about able to the features of classify a
PETs. new PET.
I want to decide I want to on which PET create metawould be most models for suitable in a given PETs.
system environment.
I want to decide I want to on the best way to contribute to implement a PET the in a given development situation. of new PETs.
I want to evaluate
PET implementations
and develop recommendations.
I want to find new use
cases to which PETs could be applied.
To evaluate the relevance of the presented in Table 3 learning goals for different roles, we conducted a survey with practitioners. In designing the survey, we first ensured that the role of each respondent was captured. Next, the statements of each separate knowledge level were presented, and the respondent was prompted to select which statement was most relevant to the task of their specific privacy role. In addition, the respondent was informed that the statements followed a hierarchical order, meaning that selecting a more advanced cognitive process included all the previous ones as relevant. For example, in the Factual Knowledge category, choosing "I want to know the various use cases for PETs" implies that "I want to know what different PETs exist" also applies.
The role-specific insights are presented in Table 4 which separates the results based on the reported role. Table 4 utilizes a heat map to illustrate the frequency by which a particular option was chosen. Thus, the number displayed in each cell represents the aggregated number of responses that the corresponding option received, considering the previously introduced hierarchical setup.
As can be seen from Table 4, roles from the three different privacy role categories possess different learning goals, which is made particularly salient by our utilization of Bloom's Revised Taxonomy. Table 4b suggests that legal experts in the privacy compliance process would be most concerned with obtaining factual knowledge about PETs. This is plausible, as legal experts would not be involved in the implementation of PETs, but rather must be knowledgeable on the topic in general, i.e., know the facts. In Table 4a, a clear preference from technical experts towards factual and procedural knowledge can be observed. Thus, these experts must not only be cognizant of the facts, but also be skilled in the procedural know-how required for the implementation of PETs. Another interesting finding arrives with an analysis of the learning goals of Go-Between roles, whose preferences seemingly reside distinctly in conceptual knowledge. Looking to Table 3 for an explanation, one can see that conceptual knowledge truly lies on the border between factual and procedural knowledge, in the way that factual knowledge becomes important more from an IT architecture and policy point of view, rather than pure implementation. Indeed, members of the Go-Between category do exist to bridge this gap, serving as a crucial link between legal mandate and technical specification.
In this research in progress, we explore the educational needs of privacy professionals with respect to learning about Privacy-Enhancing Technologies. Under the framework of Bloom's Revised Taxonomy, we subdivide PET education into learning goals based on six cognitive processes and four knowledge levels. Moreover, we probe the relevance of each of these categories with different subgroups of privacy professionals: technical and legal experts, as well as Go-Betweens. The results of the survey provide insights into differing educational needs governed by the requirements of each role.
The practical relevance of this work is grounded in the underlying complexities of state-ofthe-art PETs, which, without the necessary expertise, can hinder their adoption, calling for focused educational efforts to foster the development of such expertise. Looking forward, we plan not only to continue working on making knowledge on PETs open, accessible, and understandable, but also to do so in a way that considers the expertise of the learner. Our next steps include the creation of learning material on PETs, the validation of such material, and the deployment of an e-learning platform to encapsulate the learning content. In the creation of learning material, the findings presented in this work will be integral to tailoring the learning experience to different professional backgrounds with specific learning needs. The e-learning platform will provide the opportunity for collaboration with industry partners, further closing the gap between academia and industry on the topic of Privacy-Enhancing Technologies.
This work has been supported by the German Federal Ministry of Education and Research (BMBF) Software Campus grant LACE 01IS17049. [9] G. W. Van Blarkom, J. J. Borking, and J. G. E. Olk, Handbook of privacy and privacy-enhancing technologies, Privacy Incorporated Software Agent (PISA) Consortium, The Hague, vol. 198, p. 14, 2003. [10] Information Commissioner’s Office, Privacy-enhancing technologies (PETs), 2023. [11] The Royal Society, From privacy to partnership: the role of Privacy Enhancing Technologies in data governance and collaborative analysis, 2023. [12] United Nations, The United Nations Guide on Privacy-Enhancing Technologies for Official
Statistics, 2023. [13] V. Braun and V. Clarke, Using thematic analysis in psychology, Qualitative research in psychology, vol. 3, no. 2, pp. 77–101, 2006. [14] M. Wiesche, M. C. Jurisch, P. W. Yetton, and H. Krcmar, Grounded theory methodology in information systems research, MIS quarterly, vol. 41, no. 3, pp. 685-A9, 2017. [15] L. W. Anderson and D. R. Krathwohl, A Taxonomy for Learning, Teaching, and Assessing: A
Revision of Bloom’s Taxonomy of Educational Objectives. Longman, 2001. [16] C. Servin, C. Tang, M. Geissler, M. Stange, and C. Tucker, Enhanced Verbs for Bloom’s Taxonomy with Focus on Computing and Technical Areas, in Proceedings of the 52nd ACM Technical Symposium on Computer Science Education, Virtual Event, USA, 2021, p. 1270.