Customer experience vs data security in AI deployments in the financial industry Bachelors Thesis – Proposal Ankit Kumar Sharma Ankit.sharma@stud.hslu.ch Supervisor: Sophie Hundertmark sophie.hundertmark@hslu.ch This Bachelor Thesis Proposal was submitted as part of the requirements for the BSc in International Business Administration at the School of Business, Lucerne University of Applied Sciences and Arts. December, 2024 1 Table of Contents 1. Introduction .......................................................................................... 3 o Background and Problem Statement o Research Gap and Relevance o Research Questions and Objectives 2. Theoretical Background ...................................................................... 9 o AI in the Financial Sector o Finance Sector and Customer Experience o Legal Framework for Data Protection and Privacy o How to Prioritize between Customer Experience and Data Security 3. Methodology ....................................................................................... 13 o Research Design o Empirical Design o Data Gathering o Data Analysis o Attention to Ethics o Limitations of the Study 4. Bibliography ........................................................................................ 17 5. Initial Draft of Bachelor’s Thesis …………….................................. 20 6. Annotated Bibliography ..................................................................... 21 7. Action Plan ……………....................................................................... 27 8. Reflection Report ……………………................................................. 29 9. Declaration of Sole Authorship .......................................................... 30 2 1. INTRODUCTION 1.1 Background & Problem Statement Artificial Intelligence (AI) has become the front-and-center topic for various groups of individuals regardless of their profession, age or geographical location. In the past few decades, not only has digital technology outpaced our ability to catch up, but so has the emersion of newer technologies and tools such as AI. In order to better understand the meaning of AI, Aldoseri et al., (2024) describe artificial intelligence as an ability for systems to comprehend, develop and solve complex tasks set forth. The same article also suggests that there is no singular definition for AI and variances include the ability to utilize and analyze data without requiring to program such convoluted responses. Currently, the combination of AI and digital technologies have become essential to an extent where a company deploying such tactics benefit heavily from enhanced customer relations and experiences (Huang et al., 2021). AI has therefore grown at an alarming rate where firms with applications of AI have advanced farther than several juggernaut academic research institutions in terms of research and development (Färber & Tampakis, 2024, p.38). However, focusing on the field of finance and banking, AI has been increasingly being adopted by Swiss financial institutions, banks and insurance companies (Swiss Financial Market Supervisory Authority [FINMA], 2023). These applications usually range from chatbot integrations to claims processing and distribution, fraud detection systems and pricing models. For instance, Swiss banks like UBS are harnessing the capabilities of AI to power their chat bot systems and utilizing AI projections to grant immediate credit (Reuters, 2024). FINMA (2023) also detailed several challenges arising out of these applications which generally surround areas such as accountability, reliability, robustness, and transparency. It has become apparent that customers do tend to revel in convenience and enjoy added benefits and features of AI. Aleksandrova et al. (2023) explore how several individuals in the field of finance may view AI as a threat to their livelihood and career, however, they emphasize that AI acts as a tool rather than a replacement for employees. Institutions implementing the use of AI in financial control, insurance and risk management procedures often replace repetitive non-cognitive tasks with the use of AI while enabling human work to be much more cognitive. Financial services companies are rapidly adopting AI as the customer trends point in the direction of integration of such tools into their day to day client interactions. Aleksandrova et al. (2023) point out to how banks, insurance companies and firms in the FinTech (Financial Technology) field implement AI into their operations, services and customer support 3 departments. An article written by Chung et al., (2020) on Mckinsey & Company argue that customers have now come to expect integration of AI into their financing and banking needs as a necessity. These users tend to utilize the personalization features offered by the models, recommendations or suggestions and consistency offered by AI. For instance, AI-powered tools can seamlessly bridge the gap between switching from web browsers or similar platforms to a banking application downloaded on the user mobile for convenient transition. Alternatively, approval of loans and credits done instantly by the AI models online or assistance and support of resolving transaction issues can be facilitated by AI at the benefit of end consumers. However, the rapid growth and advancement of such applications could raise concerns regarding data protection and consumer security. In order to provide convenient services and lightning-fast response times, AI models have to be fed an incredible amount of sensitive and private user data to enable them to automatically recommend, feature useful resources or grant instant loans or credits (Ridzuan et al., 2024). Additionally, unfair practices or biases may also arise out of these solutions that may lead to unintentional incidents such as discrimination on loan approval for applicants based on genders or race or opaque decision making procedures that would be difficult to explain to end consumers. Certain other non-consumer related challenges might include the lack of expertise in data scientists that are well-equipped and experienced with sifting through AI-processed data. Sriram et al. (2023) also add to the expertise aspect of AI-powered tools by elaborating how information and data fed to these tools might be interpreted falsely or in a way that is not intentional. They discuss the use of a tool for processing insurer policy data which challenges the user with population of criteria and datasets as intended by the institution. Svetlova (2022, p. 720) illustrates that the inter-twined nature of large financial institutions and systems paired with opaque AI decision-making tools may not only lead to flawed risk assessment when it comes to granting individual loans but may also trigger similar outcomes in other firms and institutions. Simultanesouly, during a bad economic cycle, data provided to these models may interpret such cycles as an ordinary situation and may lead to negative outcomes for end consumers. This could lead to loss of trust from consumers and at the same time might risk predatory practices from less reputable institutions that target such consumers. Behavior profiling and group-based discrimination standout as one of the most prevalent forms of ‘AI attacks’ when it comes to data processing (Mühlhoff, 2023). However, data protection regulations with regards to AI data processing is yet another topic that Switzerland takes seriously. According to Butt (2024, p. 13), the GDPR (General Data Protection Regulation) of 2016 imposes several challenges for firms and institutions implementing the use of AI technology in their daily operations. For instance, GDPR requires establishments to only process necessary data useful for carrying out their operations. This might contradict the existence and impact the development of 4 AI based on behavioral learning and pattern recognition. Simultanesouly, GDPR guidelines requires the firms to provide consent from the end user to submit their data for processing. The line is thus blurred when it comes to processing large amounts of data that might be unintentionally fed by the end user to these systems which is then gathered, sorted and analyzed by the AI systems. Data retention is yet another aspect of the GDPR guidelines that restrict companies from storing data that is unnecessary after the completion of a transaction or service provision. However, for AI systems to detect fraudulent instances or to improve their algorithms for future scenarios, data processing, refining, and re-processing is a core element. This leads to conflicting and contradicting applications of AI when abiding by the GDPR guidelines. 1.2 Research Gap and relevance In a Mckinsey study by Anant et al. (2020), customers are highly distrustful of their personal data being packaged, sold and re-sold by numerous data harvesting and data capture points. Although personalization of services remains the focal point, several customers tend to distrust firms that lack transparency and act out of capitalistic endeavors. The same study cites that 87% of customers would not conduct any business with an institution or firm that misuses their data or if the customers feel unsafe sharing their data with them. This dilemma creates a natural fork between the choice of data protection or utmost convenience for customers. In such a situation, most customers tend to lean on the side of convenience and are even willing to share their information provided that they trust the firms in question. This delivers an opportunity for firms in various industries that plan on utilizing AI models while catering to client needs by ensuring data protection and garnering a perception of trustworthiness. In the book by Leenes et al. (2017), the authors explain the importance of taking into consideration consumer trust factor, data protection and privacy prior to harvesting such information. It states that failure to provide securitized and privatized data capturing techniques leads to erosion of trust amongst the customers and lack of transparency further exacerbates the situation by pushing them away. For instance, the book discusses about the implications of the Cambridge Analytica scandal where the political firm harvested countless amounts of data to create ‘personas’ that were used to influence the 2016 election and the Brexit referendum. There was an enormous commotion from the public directed towards Facebook which reduced their trust in the firm and led to several users deleting their profiles. Simultaneously, Cambridge Analytica faced equal backlash which had several legal, financial, business and industry-wide repercussions. Therefore, it is critical for firms in the financial sectors to explicitly indicate their data harvesting and processing practices prior to onboarding customers. 5 Despite all these studies and research papers, there is little concrete evidence to suggest that customers prioritize data protection over experience based on AI features that engage in mass data harvesting and profiling. Anant et al. (2020) indicate in their study that customers prefer protecting their personal information and generally permit healthcare or finance sector firms to openly utilize them however, it is not clear whether these opinions are shared in other regions of the world. Additionally, the reasoning or rationale behind the consumers to express such opinions is not characterized by age, tech-savviness, convenience, cultures, or other such attributes. The researcher conducted in-depth analysis of several articles and books published of scientific orientation and failed to retrieve excerpts detailing the logic and rationale of end consumers with regards to the topic of data protection vs. customer experience in an AI-oriented world within the financial sectors. Several studies pointed out to the rules and regulations laid down by GDPR and how AI learning and data processing algorithms requires mass data processing and re-processing which conflicts with an institution operating within EU without legal risks. However, most studies do not factor in the consumer perspective and what factors drive them towards prioritizing or selecting one or the other. This depicts the need for research from a different point of view where numerous attributes are weighted and calculated when assessing customer preferences in the financial sector. Additionally, these studies lacked emphasis on situations or scenarios relating to the prioritization of the two elements. For example, the Mckinsey study conflates healthcare and financial sectors as highly trustworthy from the consumer perspective whereas other sectors are not considered as trustworthy. Questions regarding whether situations within the financial sector with regards to loan approval, mobile banking or other such primary and secondary activities would result in similar outcomes arises. Experiences of customers with regards to predictive conversations or immediate fraud detection systems also require further research. A small pocket for behavioral economics also exists when evaluating the dilemma, a customer is involved in amongst data protection or convenience. However, the researcher will not include these dynamics for this thesis due to time constraints and scope of the paper. These gaps in available scientific materials leads the researcher to conduct thorough research and analysis of customers encompassed within the financial sector on routine basis. Additionally, quantifying and measuring these two parameters with the help of attributes relevant to the general public can provide all the stakeholders involved with insightful data and information regarding the criteria that customers deem valuable over others. Demographics often play a key role in determining motivational factors in decision-making for customers. Hence, such gaps in scientific research exist which are to be covered by this thesis. 6 1.3 Research Question and Objectives The aim of this study revolves around understanding customer preferences and what drives them towards selecting AI-driven customer experience or data protection as a top priority within the constraints of financial applications and channels. The research question answered by this thesis is: What is the focal point for customers in an AI-driven finance sector of Switzerland: customer experience or data protection? This research question is to be answered in tandem with several objectives that encapsulate the core of this research thesis. In order to address the core dilemma, the decision-making factors and utility generated out of individual experiences is to be quantified vividly. The following objectives help the researcher reach a conclusion that provides practical implications and action-oriented outcomes and scenarios that would be applicable to firms and institutions within the finance industry in Switzerland: - When engaging with AI-directed financial services and products, what key factors play a decisive role in persuading customers to opt for customer experience over data protection - Which specific scenarios illustrate an outlier when benefitting from financial services where customers might prioritize one of the parameters over the other? - Do demographics such as age, wages, gender, or education levels sway their decision or do other attributes such as tech-savviness or being an early adopter of new technology play a crucial role in shaping their decisions? - Under what circumstances are the customers open to compromising their beliefs or opinions in between the two parameters? - Do customers believe that the legal and regulatory framework together with the policies and regulations implemented by the financial institutions align with their own standards of data protection? Each of the above objectives provide detailed information and substantive reasons that help answer the research question. Additionally, the outcomes of the aforementioned objectives could shape institutional policies and regulations when it comes to data protection and customer experience curation. The results of this research thesis are expected to conclude with precise details stating the factors involved in decision-making as well as the situations that may set the parameters to flip in specific scenarios. 7 1.4 Significance of the Research This research thesis aims at contributing to numerous facets of society. From an academic standpoint, the outcomes of this thesis provide educational institutions with detailed study on how an ever-evolving world catches up to the rapid changes of technological advances all the while ensuring legal and ethical standards. AI ethics are likely to establish itself as a core branch of ethical studies where this study will contribute positively towards future research and studies. Additionally, this thesis also bridges the literature gap currently present with regards to this topic. From an industry perspective, the outcomes of this thesis aim to provide practical implications for firms and institutions in the finance sector with concrete and substantive evidence. The results arising out of this thesis are also set to garner trust among customers and institutions which would impact society in a positive manner. Additionally, following the developments from leading financial and fintech companies re-writing their framework for AI ethics and customer experience, policy makers would also benefit from the application of these measures in a business environment. This aims to persuade the policy makers to revise and re-write the ethical guidelines and regulations with respect end consumers in mind. Therefore, the researcher believes that topic allocated is of crucial importance and is set to impact social, legal, business and ethical frameworks in a positive manner. It encompasses one of the most sought-after technology and ensures that customers have deep-rooted trust with the firms in question which would be backed by substantive study. 8 2. THEORETICAL BACKGROUND 2.1 AI in Finance Industries According to Remolina and Gurrea-Martinez (2023), AI in the finance and banking sector began with the adoption of simpler programs and software that automated repetitive tasks such as cash withdrawal from an ATM or payment processing networks. This quickly took over other crucial aspects of banking industry such as chatbots, detecting and flagging fraudulent activities, and sorting, analyzing user data to provide sound recommendations to firms. Similar story plays out in adjacent sectors as well such as stocks trading and insurance. Algorithmic trading includes very simplistic AItools that follow a set of rules and instructions which have been in play since the 1990s. Gradually, this evolved to high-frequency trading devices that interpret real-time data for immediate execution as well as several portfolio management tools such as Wealthfront. AI applications within the insurance industry involved the usage of statistical data for risk management, processing of customer claims reports and eventually offering a direct point of contact in the form of chatbots. Remolina and GurreaMartinez (2023) discuss the Robo-Advisers case study from More Wealth where the firm’s implementation of AI-powered systems allowed them to reduce costs by increasing efficiencies, provided scalability, and enhanced user experience via personalized customizations. However, these applications came with their own set of challenges particularly with compliance and unbiased outcomes arising out of machine learning models. Consumer trust was seen as yet another key issue especially when the decision-making was not deemed transparent. Yilmaz et al. (2023) also expand on the previously mentioned benefits but add to portfolio management use cases for creating a much more diversified investment cluster using algorithms and data processing. In the absence of market stability, reinforcement learning has assisted certain firms to price and risk assess numerous portfolios, execute trades based on those risks analyzed and calculate tax policies to lower tax burdens to the minimum. Zheng et al. (2023, p.65) highlight the numerous challenges that AI face in FinTech, specifically transparency. They argue that the application of blockchain systems and networks are far more reliable from a customer standpoint due to their decentralized nature. However, when it comes to AI applications, the trust erodes quickly. Citing the case of Industrial and Commercial Bank of China, a ransomware attack was detected in 2023 where a single click from a customer of the bank led to large-scale data extraction of sensitive information. These challenges are exacerbated when numerous technologies and systems are in use. 9 2.2 Finance Sector and Customer Experience In a study consisting of 226 participants, Tulcanaza-Prieto et al. (2023) examine the impact AIdriven tools and services have on them in the banking sector. The outcomes indicated that consumer trust in these solutions is high provided that the institutions offering such services are trusted by the public and their data is securely managed. However, the services being offered have the obligation to work at the customer’s convenience and at their whim in order to instill a feeling of trustworthiness. Access to information at lightning fast-speed is now an expectation for customers and their willingness to adopt such technologies also resides in the fact that the users are well-educated in general and are aware of AI systems. Noreen et al. (2023) also emphasize the importance of education and general awareness about technological advancements for customers interacting with AI-powered tools and services. Certain customers inherently accept the amalgamation of AI into finance simply due to the fact that such countries, for example China, exhibit a strong digital space with advanced tools and software designed for convenience. The presence of advanced digital infrastructure automatically provides a heightened sense of data security for these customers especially when coupled with stringent regulations and legal framework surrounding data protection. On the contrary, cultural differences, such as in Saudi Arabia or Iran, illustrate a pessimistic view of AI as they are deemed to be the competitors of human work which would eventually drive them out of jobs. The capacity for AI to deliver personalized experiences that are curated to an individual’s personal preferences and desires is described as one of the most impactful factors when interacting within the finance sector (Sanodia, 2024, p.253). Moreover, other favorable sentiments expressed by users depict the responsiveness and duration required for executing tasks or solving queries. Predictive analytics that emphasize on investments or savings plan based on customer data is valued highly as well. However, the customers also share their concerns with the author where, excluding trust and data security, lack of human touch within these systems is approached as a negative characteristic of AI-driven tools. Another key challenge expressed was the familiarity and the innate ability of these services to overwhelm the consumer upon changing the layouts of the platform. Sanodia (2024, p.255) suggests that education and perhaps crash-courses might assist in resolving these minor issues. Another study also re-affirmed the findings mentioned prior regarding lack of human touch (Shaikh et al., 2024, p.177). The author concluded that despite benefitting enormously in terms of speed and convenience, the end customers preferred human interaction, implying that AI tools should be used as a complement to conventional means rather than a replacement. Trust, data security and the legal and regulatory framework involving AI ethics remained at the forefront of customer concerns as implied by numerous studies and research papers previously. 10 2.3 Legal Framework for Data Protection and Privacy The Swiss Data Protection Act (DPA) of 1992 is a legal and regulatory framework set by Switzerland to ensure compliance from all institutions and firms registered and operating in Switzerland with respect to privacy, data harvesting, and usage (Federal Office of Justice, n.d.). This framework was further revised in 2020 to keep up with advancing technologies and integrating blockchain as well as AI driven products and services. Specific data security and technical control protocols are established in the revised Federal Act on Data Protection (FADP) which further strengthens trust in these institutions for resident of Switzerland. Moreover, failure to comply with these standards, particularly in an event where data is transferred outside of the country, results with fine and penalties and in severe cases sanctions. GDPR or General Data Protection Regulation is an expansive data protection law which entered effect in May 2018 for nations situated in the European Economic Area or part of the European Union (GDPR-Info, n.d.). Principles of focus include integrity and confidentiality, storage restrictions, access only to necessary data, limited or restricted data accumulation, and accuracy of the collected data. The GDPR law grants substantial control in the hands of individuals as they command the right to data access, erasure, portability, and objection to its usage for marketing purposes. The law holds organizations and establishments accountable for data breaches where penalties may lead to fines of up to 20 million Euro or 4% of annual global revenue. Revolut is a UK-based financial technology business that offers banking services to most of Europe, currency exchange, remittances etc. (Robinson, 2019). The company’s AI enhanced tools targeted the prevention of fraud and money laundering, however, their implementation and execution of the AI led to several legal and ethical complications for the business. For example, the tool detected numerous transactions as fraudulent and flagged them in a false-positive setting. In order to vet the information and data incoming, Revolut requires the ability to process massive amounts of data which raised concerns over its compliance to the GDPR. Yahoo Finance (2023) highlights the lengthy application process and ability to acquire a UK banking license leads to wider public and legal scrutiny regarding the compliance to data protection laws in UK and EU. Although Revolut’s jurisdiction is within the UK, it is undeniable that the firm operates essentially in Europe and therefore, failure to comply by UK data protection laws raised serious questions over the ability to comply with GDPR, especially after establishing GDPR law as one of the world’s most strict data protection law (Bloomberg Law, n.d.). Such allegations dissolve the firm’s trust amongst the general public and leads to a negative perception of the use of AI in finance sector as untrustworthy. 11 Ridzuan et al. (2024) illustrate how firms struggle to balance between providing AI-driven innovative services while complying with data protection laws in their respective states. Naturally, the benefits of integrating these technologies are mandatory in a world where AI-powered firms display a distinct advantage. However, skilled teams are necessary to monitor, adjust and pivot when bias, unfairness or breach of data is detected. Ridzuan et al. (2024) also point how Europe and Switzerland are taking strides in encouraging and fostering innovation within these sectors with Basel Committee of Banking Supervision and European Insurance and Occupational Pensions Authority promoting the use of AI in an ethical context. 2.4 How to prioritize between Customer Experience and Data Security Ahmed et al. (2024) examine the influence AI has on the financial sector with respect to customer interactions. It is suggested that innovation and customer experience must remain at the forefront of development and progression of primary applications all the while ensuring that data security standards are upheld. Certain concrete measures such as two-factor authentication (2FA) as well as highly protected (via encryption) data storage systems are recommended. Following such a customercentric path, cumulative developments rather than monumental leaps in AI-driven systems is recommended. Prakarsh et al. (2024) effectively communicate the balance by suggesting the ability for users to monitor and control their data, all the while being able to ‘pull the plug’ from such services if they deem their privacy is being violated. In such a context, transparency from the firms is of critical value and assists in balancing innovation and customer experience with data protection and legal compliance. The researchers in that study illustrate the example of Clearview AI, an application for law enforcement authorities used by coupling facial recognition technology with the publicly available images of individuals on social media platforms. Although the application enhanced the performance and reduced inefficiencies for law enforcement authorities, it created a whole new array of privacy concerns including legal compliance with GDPR and unfair algorithmic biases. This case exemplifies the requirement for such firms to transparently communicate their data harvesting practices combined with legal compliance. 12 3. METHODOLOGY Methodological design of a research thesis is an integral part of understanding and rating the quality of research and analysis conducted. Substantive and concrete analysis often lead to practical implications that could be applied on mass industry wide. 3.1 Research Design For the researcher to better understand the forces driving a customer’s decision towards engaging or disengaging with technologies supportive of AI systems and tools within the financial sector, a descriptive segment of research is deployed for substantive outcomes (Erickson, 2017). These consist of closed-ended questions that target the rationale of customers in a quantifiable manner, particularly through the use of surveys or questionnaires. For instance, a typical question answered by this aspect of research include ‘What percentage of users feel the need to utilize financial services that integrate AI-powered tools?’ or ‘What demographic subsets discourage customers from sharing their private information with firms within the finance sector utilizing AI-driven systems?’. Erickson (2017) further adds the implementation of exploratory research whereby, due to lack of existing research or data, the combination or addition of open-ended questions help delve deeper into the subject matter. With respect to this research thesis, exploratory aspect of the research design will be fruitful for the researcher in narrowing down the themes, topics and factors that motivate customers to prioritize data privacy over customer experience. Cross-referencing these outcomes with those from the descriptive segment of the research, the researcher can provide concrete evidence regarding the research question. Figure 1 below portrays a decision tree for a variety of research studies based on the requirements and tasks at hand (Centre for Evidence Based Medicine, n.d.). Although this article focuses on medicinal research, the overarching ideas can be applied to traditional research. 13 Figure 1 Evidence-based research tree. Adapted from Study designs, by Centre for Evidence-Based Medicine, n.d., University of Oxford (https://www.cebm.ox.ac.uk/resources/ebm-tools/study-designs). 3.2 Empirical Design As per Davies and Hughes (2014), qualitative research revolves around problem statements arising out of theoretical frameworks and concepts which demand inductive ideology while quantitative research generally takes quantitative data to be paramount in a deductive approach. The target population is primarily customers in Switzerland benefitting from financial services or tools that integrate artificial intelligence features. The researcher categorizes the population surveyed primarily via key demographic groups, particularly the young customers – aged 18 to 30, the frequent customers – aged 31 to 50, and the experienced customers – aged 51 and above. These 3 groups could illustrate certain difference with regards to digital literacy, however, Buckingham (2020) argues that digital literacy is generally arising out of educational background, social and cultural norms, and other such factors. Another focal point for the researcher includes the types of services utilized by customers with regards to the research question and whether a specific set of services imply a different standard or perspective for end customers in their dilemma of prioritizing data protection over customer experience and innovation. All such data sets assist the researcher in generating sound 14 analysis of the problem statement at hand and further enables him to understand various customer attitudes. The researcher is set to deploy probability sampling techniques, specifically the stratified sampling. In this method of sampling, the researcher can target specific groups of the population that interact with AI-driven financial sectors. This is further divided into people of varying age groups, digital literacy, annual incomes, and similar characteristics useful for this thesis. 3.3 Data Gathering Initiating with a background knowledge and understanding of existing studies or similar studies, secondary data will be compiled and analyzed. These sources would ideally include case studies, industry analysis and scientifically acceptable academic literature studies. Sources would include the official departments and associations of Switzerland such as FINMA, Swiss Bankers Association, EU GDPR authorities, and encompass independent research firms and organizations as well such as Mckinsey or similar Swiss counterparts. Primary data sources will essentially be in the form of surveys sent mostly online or in some cases on paper. These surveys aim at gathering quantitative data in the form of Likert scale or rating questions as well a handful of qualitative data points via the addition of open-ended questions mostly requiring the participants to articulate trade-offs or opinions regarding this matter. The method of transmission is still under consideration; however, popular platforms and channels include survey monkey and Google forms. 3.4 Data Analysis For questions emphasizing the qualitative aspects of the survey, themed responses will be analyzed and structured in an organized manner through manual coding. The keywords will dictate the outcomes of these responses in a transparent manner where evident-based analysis would illustrate the reasons or factors for customer choices. With regards to quantitative analysis, simpler statistics would be an integral part of the study, such as mean, medians, and standard deviations. Likert rating scale coupled with demographic attributes assist the researcher in pin-pointing the factors and key characteristics of the participants surrounding the subject matter. Additionally, regression analysis will also be a vital component of data analysis to unravel the dependence of customer’s experience vs data protection preferences. 15 3.5 Attention to Ethics Given that this research thesis is part of an ethical study in of itself, the researcher aims to carry out strict measures of data protection and data capture via surveys distributed. Undoubtedly, the research must comply with the Swiss Data Protection act and only capture relevant data useful for this study. Storage, handling and masking the information, specifically the personal details, of participants is of utmost priority which will adhere to the standards set forth as per the laws. Consent of participants together with the assurances of their confidentiality are a given prior to engaging with them. 3.6 Limitations of the Study Understandably, this research thesis recognizes the vast amount of challenges embedded in a study of this scope and magnitude. Initially, the survey might represent some bias from participants due to certain social expectations of being open-minded and accepting towards newer technologies and innovation. To counter such scenarios, the researcher will emphasize on the anonymity and invisibility aspects of the survey. This enables the participants to express honest perspectives. Another bottleneck might include access to data from large institutions that might withhold such information for business reasons. However, the researcher aims to access a vast amount of information from public sources that often cite such sources in their studies and dossier. Time is yet another limitation in this study. Since the submission deadline of this thesis affords the researcher only 4 months’ worth of time, the scope of the study will be limited. Additionally, the study is restricted specifical to the geographical region of Switzerland, implying that the outcomes might not be applicable to other regions or nations of the world. 16 4. BIBLIOGRAPHY Ahmed, F., Hussain, A., Khan, S. N., Malik, A. H., Asim, M., Ahmad, S., & El-Affendi, M. (2024). Digital risk and financial inclusion: Balance between auxiliary innovation and protecting digital banking customers. Risks, 12(8), Article 133. https://doi.org/10.3390/risks12080133 Aleksandrova, A., Ninova, V., & Zhelev, Z. (2023). A survey on AI implementation in finance, (cyber) insurance and financial controlling. Risks, 11(5), Article 91. https://doi.org/10.3390/risks11050091 Aldoseri, A., Al-Khalifa, K. N., & Hamouda, A. M. (2024). AI-powered innovation in digital transformation: Key pillars and industry impact. Sustainability, 16(5), Article 1790. https://doi.org/10.3390/su16051790 Anant, V., Donchak, L., Kaplan, J., & Soller, H. (2020, April 27). The consumer-data opportunity and the privacy imperative. McKinsey & Company. Bloomberg Law. (n.d.). The EU’s General Data Protection Regulation (GDPR). Retrieved November 16, 2024, from https://pro.bloomberglaw.com/insights/privacy/the-eus-general-data-protectionregulation-gdpr/#what-is-the-gdpr Buckingham, D. (2020). Epilogue: Rethinking digital literacy: Media education in the age of digital capitalism. Digital Education Review, 37, 230–239. https://doi.org/10.1344/der.2020.37.230-239 Butt, J. (2024). The General Data Protection Regulation of 2016 (GDPR) meets its sibling the Artificial Intelligence Act of 2024: A power couple, or a clash of titans? Acta Universitatis Danubius. Juridica, 20(2), 7–52. Centre for Evidence-Based Medicine. (n.d.). Study designs. University of Oxford. Retrieved November 16, 2024, from https://www.cebm.ox.ac.uk/resources/ebm-tools/study-designs Chung, V., Gomes, M., Rane, S., Singh, S., & Thomas, R. (2020, October 13). Reimagining customer engagement for the AI bank of the future. McKinsey & Company. Retrieved from https://www.mckinsey.com/industries/financial-services/our-insights/reimagining-customerengagement-for-the-ai-bank-of-the-future 17 Davies, M., & Hughes, N. (2014). Doing a successful research project: Using qualitative or quantitative methods (2nd ed.). Palgrave Macmillan. Erickson, G. S. (2017). New methods of market research and analysis. Cheltenham, England; Northampton, MA: Edward Elgar Publishing. Färber, M., & Tampakis, L. (2024). Analyzing the impact of companies on AI research based on publications. Scientometrics, 129(1), 31–63. https://doi.org/10.1007/s11192-023-04867-3 GDPR-Info. (n.d.). General Data Protection Regulation (GDPR). Retrieved November 16, 2024, from https://gdpr-info.eu/ Huang, Z., Shen, Y., Li, J., Fey, M., & Brecher, C. (2021). A survey on AI-driven digital twins in Industry 4.0: Smart manufacturing and advanced robotics. Sensors, 21(19), Article 6340. https://doi.org/10.3390/s21196340 Leenes, R., van Brakel, R., Gutwirth, S., & de Hert, P. (Eds.). (2017). Data protection and privacy: The age of intelligent machines. Oxford, UK; Portland, OR: Hart Publishing. Mühlhoff, R. (2023). Predictive privacy: Collective data protection in the context of artificial intelligence and big data. Big Data & Society, 10(1). https://doi.org/10.1177/20539517231166886 Noreen, U., Shafique, A., Ahmed, Z., & Ashfaq, M. (2023). Banking 4.0: Artificial intelligence (AI) in banking industry and consumer’s perspective. Sustainability, 15(4), Article 3682. https://doi.org/10.3390/su15043682 Prakarsh, P., Mansi, M., & Vardhan, H. (2024). Balancing innovation and privacy: Assessing the legal implications of artificial intelligence in the context of privacy rights and data protection. International Journal for Multidisciplinary Research, 6(5). https://doi.org/10.36948/ijfmr.2024.v06i05.26746 Remolina, N., & Gurrea-Martinez, A. (Eds.). (2023). Artificial intelligence in finance: Challenges, opportunities and regulatory developments. Cheltenham, UK; Northampton, MA: Edward Elgar Publishing. Ridzuan, N. N., Masri, M., Anshari, M., Fitriyani, N. L., & Syafrudin, M. (2024). AI in the financial sector: The line between innovation, regulation and ethical responsibility. Information, 15(8), Article 432. https://doi.org/10.3390/info15080432 18 Robinson, H. (2019, March 1). Revolut faces FCA probe over compliance failing. The Telegraph. Retrieved November 16, 2024, from https://www.telegraph.co.uk/technology/2019/03/01/revolutfaces-fca-probe-compliance-failing/ Sanodia, G. (2024). Enhancing CRM systems with AI-driven data analytics for financial services. Turkish Journal of Computer and Mathematics Education, 15(2), 247–265. https://doi.org/10.61841/turcomat.vl5i2.14751 Shaikh, A. A., Kumar, A., Mishra, A., & Elahi, Y. A. (2024). A study of customer satisfaction in using banking services through artificial intelligence (AI) in India. Public Administration and Policy, 27(2), 167–181. https://doi.org/10.1108/PAP-05-2023-0060 Sriram, V., Fan, Z., & Liu, N. (2023). ECLIPSE: Holistic AI system for preparing insurer policy data. Risks, 11(1), Article 4. https://doi.org/10.3390/risks11010004 Svetlova, E. (2022). AI ethics and systemic risks in finance. Big Data & Society, 2(1), 713–725. https://doi.org/10.1177/20539517231166886 Swiss Financial Market Supervisory Authority (FINMA). (2023). Artificial intelligence in the Swiss financial market. Retrieved November 16, 2024, from https://www.finma.ch/en/documentation/dossier/dossier-fintech/kuenstliche-intelligenz-im-schweizerfinanzmarkt-2023/ Tulcanaza-Prieto, A. B., Cortez-Ordoñez, A., & Lee, C. W. (2023). Influence of customer perception factors on AI-enabled customer experience in the Ecuadorian banking environment. Sustainability, 15(16), Article 12441. https://doi.org/10.3390/su151612441 Yahoo Finance. (2023, March 3). Exclusive: Revolut audit queries spark skittish investors. Retrieved November 16, 2024, from https://uk.finance.yahoo.com/news/exclusive-revolut-audit-queries-skittish203742350.html Yilmaz, A. E., Rhyner, U., & Ankenbrand, T. (2023). Quantum computing and artificial intelligence in finance. https://doi.org/10.5281/zenodo.12759147 Zheng, X., Gildea, E., Chai, S., Zhang, T., & Wang, S. (2023). Data science in finance: Challenges and opportunities. AI, 5(1), 55–71. https://doi.org/10.3390/ai5010004 19 5. INITIAL DRAFT OF BACHELOR’S THESIS Table 1 illustrates the planned content for the Bachelor’s Thesis on the subject matter. Since this is an initial draft, the contents are not conclusive and are therefore subject to future changes. Table 1 Initial Draft of Bachelor’s Thesis SECTION DESCRIPTION Title Page Includes the title of the thesis, information of the student and coaches, the Bachelor’s Thesis client, location, date and name of the university. Abstract A short summary of the thesis that illustrates the topic, importance and some intriguing facts arising out of it. Table of Contents A structured and detailed overview of chapters of the thesis. Introduction Briefly introduces the paper illustrating the impact of AI in finance sector, customer experience importance and the challenges that might arise. Literature Review Scientific sources that cover similar topics are to be analyzed: - Artificial Intelligence in Business and Finance - Customer experience surrounding AI in finance - Challenges faced by firms (legal and operational) - Trade-offs Methodology Explaining the process of conducting primary research, specifically surveying the target market. Results Illustrating the findings from the research conducted. Discussions Interpreting and discussing the results with additional tools. Conclusions & Concluding the research with a sound summary and sharing practical Recommendations implications for the finance industry. Bibliography List of all sources referenced alphabetically. Appendices Addition of supportive tables, figures, notes, as well as survey distributed. Declaration of sole authorship included as well. Note: Table created by author 20 6. ANNOTATED BIBLIOGRAPHY Table 2 below indicates the current list of references and sources take under consideration for the bachelor’s thesis. Due to the nature of the project, this list is subject to future adjustments. Table 2 Annotated Bibliography # Source Title Type Keywords Relevance 1 Ahmed, F., Hussain, A., Khan, S. N., Malik, A. H., Asim, M., Ahmad, S., & El-Affendi, M. (2024). Digital risk and financial inclusion: Balance between auxiliary innovation and protecting digital banking customers. Risks, 12(8), Article 133. https://doi.org/10.3390/risks12080133 Digital risk and financial inclusion: Balance between auxiliary innovation and protecting digital banking customers Article Digital risk, financial inclusion, innovation, banking customers Useful for understanding how innovation in AI impacts financial inclusion and what risks need to be mitigated. 2 Aleksandrova, A., Ninova, V., & Zhelev, Z. (2023). A survey on AI implementation in finance, (cyber) insurance and financial controlling. Risks, 11(5), Article 91. https://doi.org/10.3390/risks11050091 A survey on AI Article implementation in finance, (cyber) insurance and financial controlling AI in finance, cyber insurance, financial controlling Provides a comprehensive overview of AI applications and its role in insurance and finance, focusing on operational improvements. 3 Aldoseri, A., Al-Khalifa, K. N., & Hamouda, A. M. (2024). AI-powered innovation in digital transformation: Key pillars and industry impact. Sustainability, 16(5), Article 1790. https://doi.org/10.3390/su16051790 AI-powered innovation in digital transformation: Key pillars and industry impact Article AI, digital transformation, sustainability, industry impact Highlights the significance of AI in advancing sustainability goals and transforming industries, useful for understanding broad applications of AI. 4 Anant, V., Donchak, L., Kaplan, J., & Soller, H. (2020, April 27). The consumer-data opportunity The consumerdata opportunity Report Consumer privacy, data analytics, Explores how to balance consumer data 21 5 and the privacy imperative. McKinsey & Company. and the privacy imperative Bloomberg Law. (n.d.). The EU General Data The EU General Data Protection Regulation (GDPR) Protection Regulation (GDPR). Retrieved November 16, 2024, from https://pro.bloomberglaw.com/insights/privacy/the- financial services use and privacy, relevant for discussing ethical AI implementations in finance. Webpage GDPR, data protection, privacy, compliance Explains GDPR regulations, essential for analyzing legal frameworks governing AI and data privacy. eus-general-data-protection-regulation-gdpr/#whatis-the-gdpr 6 Buckingham, D. (2020). Epilogue: Rethinking digital literacy: Media education in the age of digital capitalism. Digital Education Review, 37, 230–239. https://doi.org/10.1344/der.2020.37.230-239 Rethinking digital literacy: Media education in the age of digital capitalism Article Digital literacy, media education, capitalism Discusses the societal impact of AI on information access, critical for understanding the ethical dimensions of AI in education. 7 Butt, J. (2024). The General Data Protection Regulation of 2016 (GDPR) meets its sibling the Artificial Intelligence Act of 2024: A power couple, or a clash of titans? Acta Universitatis Danubius. Juridica, 20(2), 7–52. GDPR meets its sibling the Artificial Intelligence Act of 2024: A power couple, or a clash of titans? Article GDPR, AI Act, regulatory frameworks Analyzes the interplay between GDPR and AI-specific legislation, useful for framing the legal challenges of implementing AI in finance. 8 Centre for Evidence-Based Medicine. (n.d.). Study designs. University of Oxford. Retrieved November 16, 2024, from https://www.cebm.ox.ac.uk/resources/ebmtools/study-designs Study designs Webpage Study designs, research methodologies A practical resource for selecting appropriate methodologies for the thesis research. 9 Davies, M., & Hughes, N. (2014). Doing a successful research project: Using qualitative or quantitative methods (2nd ed.). Palgrave Macmillan. Doing a successful research project: Using qualitative or quantitative methods Book Research methods, qualitative research, quantitative research A detailed guide on conducting research, critical for designing and executing the thesis study. 22 10 Erickson, G. S. (2017). New methods of market research and analysis. Cheltenham, England; Northampton, MA: Edward Elgar Publishing. New methods of market research and analysis Book Market research, analysis, data interpretation Discusses innovative approaches to market research, providing techniques applicable to understanding AI’s role in customer-focused industries. 11 Färber, M., & Tampakis, L. (2024). Analyzing the impact of companies on AI research based on publications. Scientometrics, 129(1), 31–63. https://doi.org/10.1007/s11192-023-04867-3 Analyzing the impact of companies on AI research based on publications Article AI research, publication analysis, corporate influence Explores how companies contribute to AI research through academic publications, useful for understanding corporate roles in advancing AI technologies. 12 GDPR-Info. (n.d.). General Data Protection Regulation (GDPR). Retrieved November 16, 2024, from https://gdpr-info.eu/ General Data Protection Regulation (GDPR) Webpage GDPR, data protection, privacy compliance Summarizes GDPR's key provisions, providing foundational knowledge for analyzing AIrelated data protection policies. 13 Huang, Z., Shen, Y., Li, J., Fey, M., & Brecher, C. (2021). A survey on AI-driven digital twins in Industry 4.0: Smart manufacturing and advanced robotics. Sensors, 21(19), Article 6340. https://doi.org/10.3390/s21196340 A survey on AI-driven digital twins in Industry 4.0: Smart manufacturing and advanced robotics Article Digital twins, Industry 4.0, robotics, smart manufacturing Provides a comprehensive review of AIdriven digital twins, essential for exploring advanced manufacturing applications. 14 Leenes, R., van Brakel, R., Gutwirth, S., & de Hert, P. (Eds.). (2017). Data protection and privacy: The age of intelligent machines. Oxford, UK; Portland, OR: Hart Publishing. Data protection and privacy: The age of intelligent machines Book Data protection, AI privacy, intelligent machines Offers an indepth look at privacy challenges arising from AI advancements, useful for framing legal and ethical debates. 23 15 Mülhlhoff, R. (2023). Predictive privacy: Collective data protection in the context of artificial intelligence and big data. Big Data & Society, 10(1). https://doi.org/10.1177/20539517231166886 Predictive privacy: Collective data protection in the context of artificial intelligence and big data Article Predictive privacy, AI, big data, collective data protection Examines how predictive algorithms impact data privacy, contributing to discussions on ethical AI applications. 16 Noreen, U., Shafique, A., Ahmed, Z., & Ashfaq, M. (2023). Banking 4.0: Artificial intelligence (AI) in banking industry and consumers perspective. Sustainability, 15(4), Article 3682. https://doi.org/10.3390/su15043682 Banking 4.0: Artificial intelligence (AI) in banking industry and consumers perspective Article AI in banking, consumer perspective, financial technology Analyzes AI impact on banking services, focusing on consumer satisfaction and technology integration. 17 Prakarsh, P., Mansi, M., & Vardhan, H. (2024). Balancing innovation and privacy: Assessing the legal implications of artificial intelligence in the context of privacy rights and data protection. International Journal for Multidisciplinary Research, 6(5). https://doi.org/10.36948/ijfmr.2024.v06i05.26746 Balancing innovation and privacy: Assessing the legal implications of artificial intelligence in the context of privacy rights and data protection Article Privacy rights, AI regulation, data protection Explores the balance between innovation and privacy laws, essential for understanding AI’s regulatory challenges. 18 Remolina, N., & Gurrea-Martinez, A. (Eds.). (2023). Artificial intelligence in finance: Challenges, opportunities and regulatory developments. Cheltenham, UK; Northampton, MA: Edward Elgar Publishing. Artificial intelligence in finance: Challenges, opportunities and regulatory developments Book AI in finance, regulatory developments, financial innovation Provides a comprehensive examination of AI's applications and regulatory hurdles in the financial sector. 19 Ridzuan, N. N., Masri, M., Anshari, M., Fitriyani, N. L., & Syafrudin, M. (2024). AI in the financial sector: The line between innovation, regulation and ethical responsibility. Information, 15(8), Article 432. https://doi.org/10.3390/info15080432 AI in the financial sector: The line between innovation, regulation and ethical responsibility Article AI ethics, financial sector, regulation, innovation Discusses the ethical and regulatory challenges posed by AI in finance, highlighting key innovation trends. 24 20 Sanodia, G. (2024). Enhancing CRM systems with AI-driven data analytics for financial services. Turkish Journal of Computer and Mathematics Education, 15(2), 247265. https://doi.org/10.61841/turcomat.vl5i2.14751 Enhancing CRM systems with AI-driven data analytics for financial services Article CRM systems, AI analytics, financial services Explores how AI-driven CRM systems enhance customer relationships in financial services, with practical applications. 21 Shaikh, A. A., Kumar, A., Mishra, A., & Elahi, Y. A. (2024). A study of customer satisfaction in using banking services through artificial intelligence (AI) in India. Public Administration and Policy, 27(2), 167181. https://doi.org/10.1108/PAP-05-2023-0060 A study of customer satisfaction in using banking services through artificial intelligence (AI) in India Article Customer satisfaction, AI in banking, Indian banking services Explores how AI influences customer satisfaction in banking, providing valuable insights for consumerfocused AI applications. 22 Sriram, V., Fan, Z., & Liu, N. (2023). ECLIPSE: Holistic AI system for preparing insurer policy data. Risks, 11(1), Article 4. https://doi.org/10.3390/risks11010004 ECLIPSE: Holistic AI system for preparing insurer policy data Article AI systems, insurance, policy data preparation Presents a comprehensive AI-driven system for automating policy data, relevant for insurance innovations. 23 Svetlova, E. (2022). AI ethics and systemic risks in finance. Big Data & Society, 2(1), 713725. https://doi.org/10.1177/20539517231166886 AI ethics and systemic risks in finance Article AI ethics, systemic risks, financial AI systems Examines ethical challenges and systemic risks in AI financial systems, critical for ethical AI implementation. 24 Tulcanaza-Prieto, A. B., Cortez-Ordoñez, A., & Lee, C. W. (2023). Influence of customer perception factors on AI-enabled customer experience in the Ecuadorian banking environment. Sustainability, 15(16), Article 12441. https://doi.org/10.3390/su151612441 Influence of customer perception factors on AIenabled customer experience in the Ecuadorian banking environment Article AI-enabled customer experience, Ecuadorian banking, perception factors Analyzes how customer perception affects AI integration in banking services, relevant for consumer-centric AI research. 25 25 Yahoo Finance. (2023, March 3). Exclusive: Revolut audit queries spark skittish investors. Retrieved November 16, 2024, from https://uk.finance.yahoo.com/news/exclusiverevolut-audit-queries-skittish-203742350.html Exclusive: Revolut audit queries spark skittish investors Webpage Revolut, investor concerns, fintech audits Examines the impact of audit concerns on investor trust in fintech companies, with implications for regulatory AI. 26 Yilmaz, A. E., Rhyner, U., & Ankenbrand, T. (2023). Quantum computing and artificial intelligence in finance. https://doi.org/10.5281/zenodo.12759147 Quantum computing and artificial intelligence in finance Report Quantum computing, AI in finance, future technology Explores quantum computing's role in advancing AI financial applications, providing futuristic insights. 27 Zheng, X., Gildea, E., Chai, S., Zhang, T., & Wang, S. (2023). Data science in finance: Challenges and opportunities. AI, 5(1), 5571. https://doi.org/10.3390/ai5010004 Data science in finance: Challenges and opportunities Article Data science, financial applications, AI challenges Examines the intersection of data science and AI in finance, discussing opportunities and challenges in implementation. Note: Table created by author 26 7. ACTION PLAN The tables below illustrate a tentative plan of action from the author’s perspective with regards to the fulfillment of the research required. As the thesis is yet to be initiated, all information and plans below are subject to change. Table 3 Tasks Schedule Note: The author created this table 27 Table 4 Schedule Note: Table created by author 28 8. REFLECTION REPORT This bachelor thesis propsal is written within a short timeframe of only 4 weeks. The university offered us the opportunity to find our own client and decide on a topic of our major. However, being a non-native to the country and having limited contacts, I opted for the choice of topic selection from the university’s platform. The allocation of topic took place on 28th October which was followed by a few exchanges of emails with the supervisor, Ms. Hundertmark. Initially, this thesis was slated to be written in German languge, however, due to lack of German language skills, the researcher requested the supervisor to change the language to English. The topic in of itself is of great interest and thus acted as motivators towards the write-up of the thesis proposal. The first challenge I faced was structuring and outlining the proposal. Generally, few of my strengths lie in organization and structuring. However, due to the scope and sheer size of the thesis as a whole, I was peronally looking for additional advice or templates. This was quickly resolved thanks to the presentations available to us coupled with the guidelines and documents provided on the research topic assignment platform of the university. However, the second challenge was a more personal one. Time was very limited for me as I had under 4 weeks to finish the proposal since I was travelling out of country beginning of December. Although 4 weeks are sufficient in a vacuum, the fact that I had to attend school, work 3 days a week and spend time with my family meant that I was stretched thin. With a few days of sacrifices of sleep, I managed to relatively easily overcome this specific constraint. One noteworthy challenge that I faced as well involved finding the right literature. Since this topic had limited study and material on it, I was struggling to find scientific literature. However, since the topic involved analyzing the legal frameworks, non-profit organizational analysis and information provided by government bodies helped provide substantive insight into the subject matter. To conclude, these past 4 weeks allowed me to overcome some of my challenges and fears regarding the topic and the support from the supervisor also paved the way for an expedited completion of the proposal. Being able to excel at time-management and structure the thesis with great precision strengthened my resolve and enabled me to gain more confidence. This experience and the educational outcomes encourages me to look forward to tackling the bachelor’s thesis project in 2025. 29 9. DECLARATION OF SOLE AUTHORSHIP I, Ankit Kumar Sharma, hereby certify that the attached work, Customer experience vs data security in AI deployments in the financial industry, is wholly and completely my own work and that I have indicated all sources (printed, electronical, personal, etc.) that I consulted. Any sections quoted from these sources are clearly indicated in quotation marks or are otherwise so declared. I further attest that I have included acknowledgments of the name(s) of any person(s) consulted in the course of preparing this assignment, if any. Signed: Ankit Kumar Sharma ……………………………… Date: 27.11.2024 30
You can add this document to your study collection(s)
Sign in Available only to authorized usersYou can add this document to your saved list
Sign in Available only to authorized users(For complaints, use another form )