1 Abstract 1.1 Objectives Biases inherent in electronic health records (EHRs), which are often used as a data source to train medical AI models, may significantly exacerbate health inequities and challenge the adoption of ethical and responsible AI in healthcare. Biases arise from multiple sources, some of which are not as documented in the literature (e.g., bias in medical devices measurement). Biases are encoded in how the data has been collected and labeled, by implicit and unconscious biases of clinicians, or by the tools used for data processing. These biases and their encoding in healthcare records can potentially undermine the reliability of such data and bias clinical judgments and medical outcomes. Moreover, when healthcare records are used to build data-driven solutions, the biases can be further exacerbated, resulting in systems that can perpetuate biases and induce healthcare disparities. This literature scoping review aims to categorize the main sources of biases inherent in EHRs. 1.2 Methods We queried PubMed and Web of Science on January 19th, 2023, for peer-reviewed sources in English, published between 2016 and 2023, using the PRISMA approach to stepwise scoping of the literature. To select the papers that empirically analyze bias in EHR, from the initial yield of 430 papers, 27 duplicates were removed, and 403 studies were screened for eligibility. 196 articles were removed after the title and abstract screening, and 96 articles were excluded after the full-text review resulting in a final selection of 116 articles. 1.3 Results Existing studies often focus on individual biases in EHR data, but a comprehensive review categorizing these biases is largely absent. To address this gap, we propose a systematic taxonomy to classify and better understand the multiplicity of biases in EHR data. Our framework identifies six primary sources: a) bias from past clinical trials ; b) data-related biases , such as missing or incomplete information; human-related biases , including c) implicit clinician bias, d) referral and admission bias, and e) diagnosis or risk disparities bias; and f) biases in devices and algorithms. This taxonomy, illustrated in Table 1, provides a valuable tool for systematically evaluating and addressing these issues. 1.4 Conclusions Machine learning and data-driven solutions can potentially transform healthcare delivery, but not without limitations. The core inputs in the systems (data and human factors) currently contain several sources of bias that are poorly documented and analyzed for remedies. The current evidence heavily focuses on data-related biases, while other sources are less often analyzed or anecdotal. However, these different sources of bias can compound each other, leading to a cumulative effect. Therefore, to understand the issues holistically we need to explore these diverse sources of bias. While racial biases in EHR have been often documented, other sources of biases have been less frequently investigated and documented (e.g. gender-related biases, sexual orientation discrimination, socially induced biases, and implicit, often unconscious, human-related cognitive biases). Moreover, some existing studies lack concrete evidence of the effects of the bias, but rather illustrate the different prevalence of disease across groups, which does not per se prove the effect of the bias. Our review shows that data-, human- and machine biases are prevalent in healthcare and can significantly affect treatment decisions and outcomes and amplify healthcare disparities. Understanding how diverse biases affect AI systems and recommendations is critical. We recommend that researchers and medical personnel develop safeguards and adopt data-driven solutions with a “bias-in-mind” approach. More empirical evidence is needed to tease out the effects of different sources of bias on health outcomes.
The Winner of the 2024 IJRM Best Paper Award for a scientific paper published in the International Journal of Research in Marketing.
As artificial intelligence (AI) applications proliferate, their creators seemingly anticipate that users will make similar trade-offs between costs and benefits across various commercial and public applications, due to the technological similarity of the provided solutions. With a multimethod investigation, this study reveals instead that users develop idiosyncratic evaluations of benefits and costs depending on the context of AI implementation. In particular, the tensions that drive AI adoption depend on perceived personal costs and choice autonomy relative to the perceived (personal vs. societal) benefits. The tension between being served rather than exploited is lowest for public AI directed at infrastructure (cf. commercial AI), due to lower perceived costs. Surveillance AI evaluations are driven by fears beyond mere privacy breaches, which overcome the societal and safety benefits. Privacy-breaching applications are more acceptable when public entities implement them (cf. commercial). The authors provide guidelines for public policy and AI practitioners, based on how consumers trade off solutions that differ in their benefits, costs, data transparency, and privacy enhancements.
Dorotic, Matilda; Fok, Dennis, Verhoef, Peter C. & Bijmolt, Tammo H.A. (2021)
Synergistic and cannibalization effects in a partnership loyalty program
The implicit promise of a partnership in a loyalty program (LP) is that the partners will gain new customers and the LP will reinforce the loyalty to focal partners. Although customers may be encouraged to cross-purchase from partners (which may create positive synergies), they can also switch among partners without forfeiting rewards (which may lead to the cannibalization of sales among partners). To explore these cross-partner effects, we analyze the evolution of customer purchases in a partnership LP across 33 partners from 16 industry sectors. We find that cannibalizations arise more frequently than synergies among partners, contributing to a “rich-get-richer” effect for high-penetration partners; e.g., 10% increase in transactions at department stores reduce transactions at apparel partners (by .04% for new transactions and by 1.18% for recurring customers); but in turn, they attract positive synergies from apparel (.11% increase in transactions by new customers and .37% for recurring transactions).
The purpose of this paper is to introduce a new practice, i.e. tool for online fundraising in nonprofit organizations, based on the assessment of Internet-induced local civic mindedness (INLCM) as a segmentation approach. We suggest that this novel approach to fundraising segmentation can be performed as the extension to the already existing analyses, utilizing the demographic and psychographic profiles. Based on nationally representative survey results of Croatian households, we develop an analytical procedure. It might help managers of community nonprofit organizations to target the relevant individuals by applying Internet marketing tools (such as Google Analytics) and approaches. Our analysis allows prediction of how likely it is that an individual (based on his or her demographic/psychographic profile) could be identified as INLCM (i.e. successfully targeted for fundraising efforts). Empirical results from Croatia demonstrate that such individuals do not have a lengthy Internet experience and do not use it frequently. The reported level of community belonging for these prospects is rather high, while their relevant Internet activities are related to maintaining the existing social ties and obtaining information about local organizations. It is interesting that they might be living in multi-generational households, either without their own underage children, or a single child.
Dorotic, Matilda & Vossen, Alexander (2014)
With a little help from your Customers
BI Marketing Magazine,
Dorotic, Matilda; Verhoef, Peter C., Fok, Dennis & Bijmolt, Tammo H.A. (2014)
Reward redemption effects in a loyalty program when customers choose how much and when to redeem
The redemption of loyalty program (LP) rewards has an important impact on LP members’ behavior, particularly on purchase behavior before and after redeeming a reward. However, little is known about the interplay between members’ purchase and redemption behavior when members are not pressured with points expiration and they choose for themselves when and how much to redeem. In this context, the effects of redemption are not straightforward, as little additional effort is required from an LP member to obtain the reward. Analyzing the behavior of 3,094 members in such an LP, we find that the mere decision to redeem a reward significantly enhances purchase behavior before and after the redemption event, even when members redeem just a fraction of their accumulated points. Conceptually, we refer to this enhancement as the redemption momentum, which is an alternative and novel explanation of the existence of pre-reward effects that does not depend on points-pressure. In addition to the overall impact of redemption on purchases, prior purchase behavior also enhances redemption decisions. Finally, we find a number of moderating effects on purchase and redemption behavior that derive from the length of LP membership, age, income and direct mailings. Our study’s most important managerial implication is that firms can avoid imposing binding thresholds and/or points expiry and still enhance members’ purchase behavior.
Breugelmans, Els; Bijmolt, Tammo H.A., Zhang, Jie, Basso, Leonardo J., Dorotic, Matilda, Kopalle, Praveen, Minnema, Alec, Mijnlieff, Willem Jan & Wünderlich, Nancy V. (2014)
Advancing Research on Loyalty Programs: A Future Research Agenda
Despite the growing literature on loyalty program (LP) research, many questions remain underexplored. Driven by advancements in information technology, marketing analytics, and consumer interface platforms (e.g., mobile devices), there have been many recent developments in LP practices around the world. They impose new challenges and create exciting opportunities for future LP research. The main objective of this paper is to identify missing links in the literature and to craft a future research agenda to advance LP research and practice. Our discussion focuses on three key areas: (1) LP designs, (2) Assessment of LP performance, and (3) Emerging trends and the impact of new technologies. We highlight several gaps in the literature and outline research opportunities in each area.
Keywords: Loyalty program design, partnership loyalty program, performance assessment, effects of strategic behavior, customer relationship management
Dorotic, Matilda & Olsen, Line Lervik (2013)
Hvordan kan bedrifter gjøre best nytte av kundelojalitetsprogrammer? :
Magma forskning og viten, 16(4) , s. 50-59.
Dorotic, Matilda & Olsen, Line Lervik (2013)
Bonus som gir økt business
Forskning.no,
Dorotic, Matilda & Olsen, Line Lervik (2013)
Fire bonusfeller
BI Business Review,
French, Sally & Dorotic, Matilda (2019)
Wirecutter/New York Times: "Romance, Heartbreak, and the Southwest Companion Pass", Money - Making Credit Work For You. link: https://thewirecutter.com/money/credit-cards/rewards-mania-companion-pass/
[Journal]
Pedersen, Pia Beate & Dorotic, Matilda (2014)
En kunst å skape lojale kunder
[Journal]
Do, Tuan Viet; Dorotic, Matilda & Bigman, Yochanan E. (2024)
Impact of moral judgments on the permissibility of high-risk AI and effectiveness of privacy-protection solutions
[Lecture]. Event
Dorotic, Matilda & Bigman, Yochanan E. (2024)
BENEVOLENCE OR BUST: HOW INFERRED MOTIVES DRIVE PERMISSIBLITY FOR HIGH-RISK AI
[Conference Poster]. Event
Dorotic, Matilda (2024)
Impact of Artificial Intelligence on Society: Citizens’ Responses to High-risk AI Applications
[Conference Lecture]. Event
Dorotic, Matilda (2024)
Impact of Artificial Intelligence on Society: Citizens’ Responses to High-risk AI applications
[Conference Lecture]. Event
Dorotic, Matilda & Velasco, Carlos (2024)
SmartFood report: Measuring impacts, scaling-up and drawing lessons learnt for Cities of the Future
[Report Research]. SmartFood Consortium 2022-2024
Dorotic, Matilda (2024)
Artificial Intelligence Policies
[Lecture]. Event
Stagno, Emanuela; Dorotic, Matilda & Doorn, Jenny van (2024)
Who Will Help? The Effect of Automated Social Presence on Individuals’ Likelihood to Act Prosocially
[Conference Lecture]. Event
Dorotic, Matilda (2024)
“Engaging Citizens in Food Diversity in Cities” Talk: Impact of Growing Plants on Green Food Consumption Attitudes and Behaviors
[Conference Lecture]. Event
Effects of nudging and hydroponic growing of plants on behaviors and attitudes towards sustainability and green food consumption - Matilda Dorotic
Dorotic, Matilda (2023)
Association for Consumer Research workshop on Consumer Privacy
[Conference Lecture]. Event
Dorotic, Matilda (2023)
Paradoxes and biases in the uptake of artificial intelligence
[Conference Lecture]. Event
Dorotic, Matilda (2023)
Roundtable on AI & Privacy - Multiple stakeholders perspectives
[Conference Lecture]. Event
Dorotic, Matilda & Stagno, Emanuela (2023)
AI in Public: The Effects of Technology Bias, Fears of Public Surveillance, and Moral Tradeoffs on Privacy Concerns
[Conference Lecture]. Event
Dorotic, Matilda & Johnsen, Jan William (2023)
Seksuell utnyttelse av barn over internett: Rapport om analyse av teknologiske faktorer som påvirker produksjon og deling av materiale som seksuelt utnytter barn over internett
[Report Research]. Norges teknisk-naturvitenskapelige universitet
Teknologiske utviklinger, spesielt de som er relatert til elektroniske tjenester på internett, gir viktige fordeler i å fremme kommunikasjon, tilgang til og deling av informasjon. Men denne utviklingen byr også på betydelige utfordringer for å gjøre internettbaserte miljøer trygge for barn og samtidig beskytte personvernet og ytringsfriheten. Denne rapporten tar sikte på å belyse den komplekse rollen teknologien og internettet har for å produsere og dele materiale knyttet til seksuell utnyttelse av barn, men også deres kritiske rolle i å skape en motreaksjon med det formål å oppdage og forhindre misbruk.
Nittitallets Amerikanske posttjenester hadde en enorm innsats for å bekjempe distribusjonen av materiale knyttet til seksuell utnyttelse av barn (CSAM). Overvåkingen av posttjenestene var vellykket mht. reduksjonen i delingen av CSAM [1], [2]. Delingen av CSAM har imidlertid eksplodert over hele verden med utviklingen og utbredt bruk av internett og elektroniske tjenester [3], [4]. I perioden mellom 2005 og 2020 har det vært en kontinuerlig økning i antall rapportert overgrepsmateriale til det globale rapporteringssenteret National Center for Missing and Exploited Children (NCMEC). NCMEC rapporterte om en økning på 35% av nytt overgrepsmateriell i omløp mellom 2020 og 2021 [5].
Den store tilgjengeligheten av elektroniske tjenester og plattformer (både på internett og mobil) gjør det mulig for lovbrytere å enkelt kontakte flere hundre mindreårige samtidig, samt gjennomføre kjøp, salg og utveksling av overgrepsmateriale med både mindreårige ofre og likesinnede. I tillegg gjør den økende utvikling av anonymiseringsteknikker (som ende-til-ende kryptering eller mørke nett-tjenester) og kunstig intelligens skapte bilder det vanskeligere å oppdage og fjerne overgrep og CSAM. Det store volumet av eksisterende overgrepsmateriale på nettet samt hastigheten nytt innhold skapes gjør arbeidet med å identifisere eksisterende og nytt materiell krevende. Manuell gjennomgang blir umulig. Teknologiske løsninger og tett privat-offentlig samarbeid mellom elektroniske tjenesteleverandører og andre interessenter (som foreldre, sivilsamfunn, myndigheter og politi) er nødvendig og alle har en rolle for både å hjelpe og bekjempe problemet.
Oppsporing og forebygging av seksuell utnyttelse av barn over internett er et nyansert og komplekst fenomen. På grunn av den komplekse strukturen som gjør produksjon og distribusjon av overgrepsmateriale enklere og anonymt, så kan politimyndigheter ikke nærme seg disse problemene alene eller isolert fra andre. Selv om alle er enige om at dette representerer et viktig sosialt problem, påvirker ofte oppnåelsen av målene til en interessent (f.eks. tillate økt overvåking av internett fra politiets side) direkte rettighetene og evnene til andre interessenter (f.eks. inntrenging i personvernet til enkeltpersoner eller brudd på kundeforhold til elektroniske tjenesteleverandører).
Det er en meningskonflikt om hvordan man skal nærme seg løsningene mellom de som støtter sterkere statlig overvåkning og de som motsetter seg det. Denne rapporten tar sikte på å fremheve de mest fremtredende nåværende og fremtidige teknologitrender, mottiltak og strukturelle og juridiske problemstillinger knyttet til fenomenet teknologiassistert overgrep mot barn. Vi håper dette arbeidet vil bidra til en bedre forståelse av kompleksiteten rundt problemet og bidra til å skape et enkelt rammeverk for et tryggere og bedre internett for barn!
Dorotic, Matilda & Johnsen, Jan William (2023)
Child Sexual Abuse on the Internet: Report on the analysis of technological factors that affect the creation and sharing of child sexual abuse material on the Internet
[Report Research]. Handelshøyskolen BI
Child Sexual Abuse on the Internet: Report on the analysis of technological factors that affect the creation
and sharing of child sexual abuse material on the Internet.
Authors: Matilda Dorotic and Jan William Johnsen
Project administered by NTNU (project manager Basel Katt)
Grant: SOBI Del III: Kartlegging og analyse av arenaer som brukes til tilgang og deling av overgrepsmateriale; Ministry of Justice and Public Security, Norway.
Dorotic, Matilda & Stagno, Emanuela (2023)
Artificial Intelligence Liabilities in Public: Privacy-enhancing and Transparency Effects
Citizen’s Perceptions of Anonymization and Transparency in Privacy-Enhancing Technology
[Lecture]. Event
Dorotic, Matilda; Do, Tuan Viet & Souza, Carlos Eduardo Caldas de (2021)
AI and Citizens' Wellbeing: Perceptions, Paradoxes and Trade-Offs For the Future of Policing
[Lecture]. Event
Dorotic, Matilda (2021)
7th EDEN Conference on Data Protection in Law Enforcement: Human After All: Data Protection in Policing, Rome, 18-19 October 2021
[Conference Lecture]. Event
Dorotic, Matilda (2021)
Workshop on Benefits and Fears of Policing in Smart City. Focus groups data collection with seven focus groups, 47 participants. Part of the Policing in Smart City Conference, Oslo, November 2021
[Lecture]. Event
Dorotic, Matilda (2021)
PRIVACY PARADOXES AND CITIZEN WELLBEING IN THE AI-ENABLED FUTURE
The omnipresence of loyalty programs (LP) across markets shows that LPs have been one of the most prominent business trends of the last two decades. Besides their traditional stronghold among airlines and grocery retailers, loyalty reward schemes have spread among nonprofit organizations like museums, charities and sport clubs, among online and offline services, and even among utility providers and business-to-business markets. On average, two-thirds of Europeans belong to at least one LP. In the UK LP penetration reached 90 % and even 94 % in Finland, according to a worldwide Nielsen study in 2016. The 2017 US census by Colloquy reports 3.8 billion LP memberships, with the strongest penetration in the retailing sector to which more than 1.6 billion memberships belong. Beyond reinforcing customer loyalty and retention, LPs can help a retailer increase its share in a customer’s wallet and to cross-sell and up-sell additional products to customers. However, as the number of companies offering LPs soar, the battle for a “place in the consumer wallet” is intensifying, resulting in the fact that more than half of all the memberships that customers sign up for are eventually abandoned, according to Colloquy’s reports. This tendency, coupled with increases in investment costs necessary to leverage benefits from LPs, makes some managers question whether supposed gains from LPs are sustainable. These managers wonder whether investments in LPs should rather be replaced with new musthaves such as mobile marketing, gamification and socialmedia leverage.
Dorotic, Matilda (2017)
Big Data and Social Media for Managers: How to Analyze Their True Impact on Profitability, Leaders Toolbox lecture
[Lecture]. Event
Dorotic, Matilda (2017)
Unlocked Versus Locked-In Customers’ Value In Contractual Setting. EMAC 2017, by Matilda Dorotic; Evert De Haan; Socrates Mokkas
[Conference Lecture]. Event
Dorotic, Matilda (2017)
It Pays to Pay Smart: Customer benefits of introducing Cashless App Payments; by Matilda Dorotic; Koen Pauwels Special Session Paper: Customer Management: From Pay & Satisfaction To Retention & Recommendation
[Conference Lecture]. Event
Dorotic, Matilda (2014)
Customers who redeem their rewards in loyalty programs buy more
[Popular Science Article]. ScienceNordic.com,
Dorotic, Matilda (2014)
The Power of New Insights
[Lecture]. Event
Dorotic, Matilda (2014)
Reward Redemption Effects in a Loyalty Program When Customers Choose When and How Much To Redeem
[Conference Lecture]. Event
Dorotic, Matilda (2014)
Symposium, University of Pennsylvania, Wharton Customer Analytics Initiative
