DATA AND AI ARE CHANGING THE WAY ORGANIZATIONS THINK, DECIDE, AND ORGANIZE. IT’S TIME HUMANITIES, MANAGEMENT AND SOCIAL SCIENCES GET INVOLVED.
Metalab

NEWS

Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua. Ut enim ad minim veniam, quis nostrud exercitation ullamco laboris nisi ut aliquip ex ea commodo consequat. Duis aute irure dolor in reprehenderit in voluptate velit esse cillum dolore eu fugiat nulla pariatur. Excepteur sint occaecat cupidatat non proident, sunt in culpa qui officia deserunt mollit anim id est laborum.

EVENTS

Explore our module designed to guide you in fully integrating digital into your strategic thinking. This program will help you craft a value proposition centered on the customer experience and implement a tailored marketing plan. You will then be equipped to optimize and enhance your digital strategy across both B2B and B2C markets.

Format:

6 days - on site (Paris) / in French

Read more...

The Executive Certificate in Marketing & Digital Management is an intensive 31-day program focused on developing high-level digital marketing competencies for senior executives, managers, and leaders. It is designed to prepare participants for significant professional growth and leadership roles on both a national and international scale.

Key Highlights:

  • Intensive Curriculum: 31 days of strategic training spread over several months to accommodate professional schedules.
  • Target Audience: Senior Executives, Marketing Directors, and Business Leaders.
  • Global Perspective: A curriculum built to address the complexities of international markets and digital ecosystems.
  • Professional Evolution: Structured to facilitate career transitions or advancement into C-suite marketing roles.

Recognition:

Certification registered with the National Directory of Professional Certifications (RNCP).

Format:

31 days - on site (Paris) / in French

Read more...

One or two-year program based in Paris and Singapore that is fully hybrid between Data Sciences, Business Analytics & Digital Strategy.

50% of students have a background in Engineering, Mathematics or Computer Science and 50% have studied Economics or Management before. Also, 50% have some professional experience. Students can take up to 18 advanced elective courses without limitations.

Ranked #3 worldwide and #1 in Europe on average over the last four years by QS, the Master in Data Sciences & Business Analytics combines faculty resources of two of the most prestigious Business and Engineering Schools in Europe ESSEC Business School and CentraleSupélec.


Recognition:

Master 1 and Master 2 recognized by the French State. 120 ECTS credits.

Read more...

[ESSEC KNOWLEDGE] by Andras Fulop and Laurence Daures-Lescourret - Professor and Associate Professor of Finance at ESSEC Business School.

Information technology has been shaping consumer credit risk for decades. The Fair Isaac Corporation (FICO) introduced FICO scores in 1989, marking a milestone in moving the credit evaluation process from humans towards algorithms. The FICO score is a credit score that takes into account five areas to determine creditworthiness: payment history, current level of indebtedness, types of credit used, length of credit history, and new credit accounts. Today, it is used in more than 90% of the credit decisions made in the U.S.

With the increasing digitalization of society in the last decade, there has been an explosion both in the collection of personal data and in the sophistication of the algorithms and computing capacity to process all this information. This clearly holds the potential to fundamentally impact the process of evaluating individuals’ creditworthiness. In this piece, we summarize some of the lessons that can be gleaned from the recent academic literature arising from the application of powerful artificial intelligence (AI) techniques to consumer credit risk.

How can AI help?

The general finding is that newer AI tools indeed live up to their promise of improving the technology of credit screening. Traditional credit scoring models such as the one behind the FICO score are based on hard information sources from individuals’ financial accounts. A relatively small number of predictive variables are usually then combined using rating scorecards, or linear models such as logit regressions. Newer AI approaches go beyond such tools in at least two important aspects. First, modern machine learning algorithms, such as tree-based methods or neural nets, allow flexible non-linear predictive relationships between predictors and the individuals’ credit risk, while tackling in-sample overfitting through sophisticated regularization methods. The overall message from the literature is that machine learning techniques tend to outperform traditional linear models, especially within higher risk groups. For instance, Walther et al. (2020), using a dataset of millions of US mortgages, showed that tree-based methods significantly outperform logit techniques. Albanesi and Vamossy (2019) compared several machine learning algorithms on data from the Experian credit bureau and found that their ensemble technique combining neural networks and gradient-boosted trees improves upon traditional models and the improvement is especially pronounced for consumers with low credit scores. Second, digitalization and AI algorithms allow the use of new types of data. Using a dataset covering around 270,000 purchases at a German e-commerce company, Berg et al. (2020) analysed the predictive power of digital footprints (the information left behind by individuals while visiting a website, such as the device, the email address, the shopping hour, etc.). They found that the accuracy of a credit risk model built on digital footprint variables is comparable and complementary to that of traditional credit risk scores. An important implication of this result is that digital footprints can help in screening some borrowers with little credit history. Another source of information that AI algorithms can benefit from is unstructured user data in the form of text or images (for instance, information from social networks such as LinkedIn, Twitter or Facebook). Using data from Prosper, a crowdfunding platform, Netzer et al. (2019) find that supplementing the financial and demographic information with the textual information submitted by prospective borrowers substantially improves default prediction. Using a similar dataset, Iyer et al. (2016) found that the market interest rate that mirrors the information available to lenders on Prosper is a more precise predictor of defaults than traditional credit scores.

What are the implications of these new techniques on consumer welfare?

On a positive note, improvements in screening technology that are particularly pronounced among riskier groups and people with scant credit history [Berg et al (2020)] can decrease asymmetric information problems between borrowers and lenders and lead to increased access to credit. This feature can be especially useful in emerging countries with limited reach of the formal banking sector and hence a lack of traditional credit information for most consumers. A further potential benefit of turning credit decisions over to algorithms is that human biases of loan officers – such as racism and in-group/out-group bias – can be short-circuited, leading to less discrimination. To examine this issue empirically, one first needs a precise definition of discrimination. Bartlett et al (2019) suggested using the interpretation of US courts, whereby any differential impact of the treatment of minority groups not related to ‘legitimate-business-necessity’ is deemed discriminatory. In the credit context, ‘legitimate-business-necessity’ essentially means variables that help in predicting default risk. Hence, to measure discrimination against minority groups, one would need to compare consumers from minority groups with peers from majority groups and the same credit risk. Given that empirically the credit risk of individuals is observed only imperfectly, this leads to an omitted variable problem. Bartlett et al. (2019) used an identification afforded by the pricing of mortgage credit risk by government-sponsored entities (GSEs) (Fannie Mae and Freddie Mac) in the US to deal with this issue. In particular, these GSEs use a predetermined grid that prices credit risk across loan-to-value and credit-score buckets. Given that the credit risk of conforming mortgages are insured by these GSEs, any access or price differences for borrowers within the same bucket are unrelated to creditworthiness. and fail to qualify as ‘legitimate business necessities’ but qualify as discrimination. Using this empirical strategy, Bartlett et al. (2019) found that FinTech algorithms also discriminate, but 40% less than face-toface lenders in pricing mortgages. What’s more, FinTechs do not discriminate in loan approval, while face-to-face lenders do.

New challenges of AI algorithms

First, more precise screening algorithms tend to lead to more inequality in the cost of credit among consumers. This increase in dispersion is particularly pronounced among minorities and riskier borrowers [see Walther et al (2020)]. Shaping policies that improve the terms of credit for disadvantaged households in the presence of such improved screening technology is a crucial topic for the regulatory debate going forward. Second, as Bartlett et al. (2019) pointed out, AI algorithms may also increase the performance of screening of consumers in non-‘legitimate-business-necessity’ dimensions. In particular, if a lender uses such algorithms to maximize profits unrelated to screening for credit risk and such profit-maximizing screening has a differential impact on protected minority groups, the company risks coming under the purview of anti-discriminatory legislation even if there is no personal bias against minorities in the algorithm. Further, the black-box nature of most AI algorithms increases the risk of such scenarios, as its functioning may not be clear to the humans operating it. Hence, a key challenge is the development of nondiscriminatory AI algorithms. The future of AI algorithms in credit decisions is bright, but its human operators must be careful to ensure that they understand how it works and aim to reduce the risk of inequitable decisions to provide fair, accurate evaluations to all.

More ESSEC Knowledge articles here

[ESSEC KNOWLEDGE] by Isabelle Comyn-Wattiau - Professor of Information Systems at ESSEC Business School.

Artificial intelligence has taken the business world by storm, bringing changes in decision making supports. It is a new addition to the information component toolkit that all organizations must implement (or at least consider). The introduction of the General Data Protection Regulation (GDPR) in Europe in 2018 and its equivalents in other countries and continents have highlighted the challenge that proper information governance poses to organizations. In this article, we tackle two elements of the new situation: artificial intelligence and information governance. After a refresher on what information governance entails, we will address two questions: how must information governance take into account the new artificial intelligence tools, which are both consumers and producers of information that needs to be governed? How can these same tools help organizations with data and information governance?

What is information governance?

Information governance refers to the establishment, in a company or organization, of an information m a n a g e m e n t p o l i c y, w i t h defined objectives and a planned implementation, including human, organizational, and technological resources. The terms “data governance” and “information governance” are used more or less interchangeably to refer to this kind of information asset management strategy. Using the term “data governance” emphasizes the digital aspect of the strategy, but may limit the scope to structured data. Information governance encapsulates all forms of data, regardless of their form: structured data in tables, semistructured data in documents, or less structured data in messaging software or images. Ultimately, we can define information governance by its objective: maximizing the value of information while minimizing the associated costs and risks. Information governance includes a set of key processes, such as managing data quality and security (availability, integrity, confidentiality, traceability). While these ARTIFICIAL INTELLIGENCE AND INFORMATION GOVERNANCE: A PROMISING PAIR? processes often use IT processes, they are often the responsibility of other business units, and rightfully so. Information governance is relevant for all areas and sectors of activity. It is especially pertinent when the three complementary dimensions of value, cost, and risk are significant. For example, take the banking sector, where the cost of information management can represent over 10% of revenues and the associated risk is significant, as it can involve, for example, handling sensitive information like money in electronic form! In the health sector, the risk – meaning information disclosure or error – is the most important feature, though we cannot ignore the other dimensions. Information governance is increasingly better organized in public and private companies, and more broadly, in all organizations. The latter are becoming more advanced, as indicated in other studies. That said, they must constantly face new challenges, including artificial intelligence.

The singular experience of governing artificial intelligence

Artificial intelligence tools store data and information to understand, decide, and learn. Data and information are increasingly voluminous and varied. As with all the company’s information, they must be under control, which is one of the key aspects of information governance. Be it data warehouses, data lakes, or another way to store information, nothing must be left off of the information map and all must fit into the dimensions of valuation, risk management, and cost control. Data quality must live up to its name: it must facilitate the use of artificial intelligence by rendering the processes of collecting, locating, and cleaning data readily available. Above and beyond these classic aspects of information governance, artificial intelligence raises new questions, like the presence of human biases in both algorithms and data and how this could lead to poor decisions. In addition to collecting personal data, and more generally, sensitive data, the use of information gathered by artificial intelligence must also include an explanation of these algorithms. This is the challenge of explainable AI, which refers to the methods that equip the “black boxes” of AI with modules that at least partially explain the results they provide. In many cases, this becomes a regulatory challenge. It can also be a trust issue for users to rely on the precious aid AI can provide without reluctance. Finally, it is an ethical challenge for organizations, who are thus able to ensure that their processes are transparent, even when they use artificial intelligence tools.

Artificial intelligence at the service of information governance

Conversely, artificial intelligence tools could also reveal themselves to be a precious tool for information governance. Therefore, mapping information must, among other things, identify personal data to ensure compliance with the GDPR. Artificial intelligence could help maintain this map, a tedious but important task, which requires analysis and decision-making skills. The volume and omnipresence of data makes this task impossible to accomplish manually. Humans will remain the masters of the information governance ecosystem, but artificial intelligence and other emerging technologies, such as blockchain, will play a very important role in helping analyze information and ensure compliance. Automation could also help with archiving and destroying information, the ultimate phase of the information life cycle, thus addressing both regulatory and efficiency needs. Software editors who specialize in information governance have reported integrating artificial intelligence into their component offering, though it remains difficult to evaluate the scope of this integration. Some use intelligence techniques to identify new data, uncover relationships in the data, and classify documents. Learning techniques make it possible to describe the type of content that the software can recognize after sufficient training. In natural language processing, a technique of artificial intelligence, the automatic recognition of characters is increasingly improving. Software robots can execute repetitive tasks, like locating bills in a set of administrative documents and providing the human with an analysis that he or she only needs to confirm, like how ATMs can process cheques automatically. Explainable AI would also help the organization demonstrate its strong information management when dealing with regulators or auditors. Finally, organizations must face regulations that are constantly changing and that vary from one country to another, in a world with increasingly rich and varied data: in this context, information governance has much to gain from making use of artificial intelligence tools.

More ESSEC Knowledge articles here

[ESSEC KNOWLEDGE] by Arnaud de Bruyn - Professor of marketing Analytics at ESSEC Business School.

The rise of artificial intelligence has naturally seen people applying it (or attempting to apply it) in countless ways, with varying degrees of success. While artificial intelligence can be a powerful tool- in the right hands, in the right situation- it is not as easy as implementing a system and tapping a few keys. This is perhaps especially true for environments that deal with human behavior, like marketing. As the saying goes, “with great power, comes great responsibility”: and marketing managers must be aware of its potential pitfalls to avoid problems. Equally important is the need to know how to properly deploy their AI tools to avoid squandering both its potential and their company efforts and resources. By understanding AI’s pitfalls, marketing managers can make the most of its opportunities. Arnaud de Bruyn and his colleagues Vijay Viswanathan (Northwestern University) Yean Shan Beh (The University of Auckland; Xiamen University Malaysia), Jürgen Kai-Uwe Brock (Fujitsu Global), and Florian von Wangenheim (ETH Zurich) explored this in their recent article. So far, AI’s biggest advancements in the business world have been related to deep learning, referring to complex, multilayered (i.e., deep) neural networks, solving difficult problems with predictive analytics. The more layers in a neural network, the more complex it is, and more “layered” networks can identify and learn more complex relationships between variables. This means artificial intelligence can learn to uncover relationships that existing statistical techniques cannot detect and that it can learn to do so autonomously. This is the main selling point of contemporary AI algorithms. While the ability of AI algorithms to autonomously create models is its strength, it is not without its challenges when it comes to putting it in action. These challenges are: a lack of common sense, objective functions, a safe and realistic learning environment, biased algorithms, understandable and controllable artificial intelligence, the paradox of automation, and knowledge transfer.

Lack of common sense

What do we mean by a lack of common sense? It is not an insult to its programmers or to those operating it; no, we mean that the algorithm itself lacks what we humans call “common sense”. We know that emotional intelligence is important, and indeed AI systems are increasingly able to recognize people’s emotions, through image recognition, voice analysis, or text analysis. But recognizing emotions is a far cry from understanding and feeling them. An AI system could learn that the words “queen” and “crown” are linked, and could even use them appropriately in a sentence, but the meaning of the words and sentences would be lost on it. Anything they have approaching common sense must be programmed into them by a person, which becomes a problem when it comes to objective functions.

Objective functions

An objective function is one that specifies the result that the AI algorithm aims to optimize (Sutton and Barto, 2018). In a marketing context, this could look like profit maximization or customer retention. The “freedom” of AI from common sense hinders its ability to define an objective function. It might be that humans understand something implicitly, but then have a hard time translating this for the algorithm. This might go awry: an autonomous car directed to “get to the airport ASAP!” might get there in record time, but having mowed down pedestrians and sped through red lights on its way. While the previous example is obviously extreme, we have already seen consequences of this play out in real life, with gender- or racially-biased systems. An outcome like profit maximization cannot be considered without allowing for the legal, moral, and ethical implications, which marketing stakeholders need to keep in mind when building and implementing their systems.

Safe and realistic learning environment

As you can imagine, all this is easier said than done. Knowledge transfer from the expert to the algorithm and vice versa is one of the biggest problems facing AI today, and the potential for costly mistakes is enormous. To avoid the fallout, it is important for AI algorithms to learn in a safe, realistic environment. Safe, in that if they do make mistakes, there is less impact on the business, and they avoid the marketing equivalent of running a red light. Realistic, in that the data resembles what they would receive in a real-life situation. This presents a challenge in marketing, because customers can be unpredictable, and a new factor (like, say, COVID-19) can throw a wrench into the best-laid marketing campaigns. While it might be tempting to think that AI reduces or even eliminates our need to understand customer behavior, it is the opposite: we need detailed customer behavior theory more than ever, as this will help us better configure our AI algorithms.      

Biased algorithms

This brings us to another limitation to AI’s use in marketing: its potential to be biased. Of course, the algorithm itself is not prejudiced, but if it is powerful enough, it could identify a characteristic like race or gender on its own and make biased predictions. How so? It might pick up on other information that acts as a proxy to the factor in question, like education or income, thereby unintentionally replicating the biases that are found in the data. In a marketing context, this could lead to outcomes like a price-optimization algorithm that aims to charge women more or an advertising algorithm that targets a vulnerable population. This has legal implications as well as the obvious ethical ones. Complicating the problem is the fact that adding the sociodemographic variable in question to the model in an attempt to clarify it could just make it easier for the algorithm to make prejudiced predictions. If marketing stakeholders do not properly understand the algorithms they are using, they might not know to challenge these troubling predictions.

Understandable artificial intelligence

The ability to understand and explain the model is another factor in the uptake of AI. If you are going to use an AI model, you need to understand why it makes the predictions it does, and to be able to interpret what the model is doing. More specifically, an AI’s human “handlers” need to be able to explain: 1) the purpose of the model, 2) the data it is using, and 3) how the inputs relate to the outputs. By understanding this, it is also possible to know why the AI system is preferable to a non-AI system.

Controllable artificial intelligence

Using the term “handlers” above was intentional: an AI system must be able to be controlled and overridden. This might conjure up images of I, Robot and killer robots, and while the reality is rather less lethal, it is still serious. One recent example is that Uber’s pricing algorithm responded to the crush of people fleeing the scene of the June 2017 terrorist attack in London by adapting (read: increasing) the ride prices to more than double the typical fare. Anyone who has taken Uber is unfortunately familiar with their surge pricing system, but in the aftermath of a terrorist attack, it made Uber seem like ruthless profiteers. However, Uber’s monitoring system quickly flagged the problem, and they had mechanisms established that allowed them to override the algorithm within minutes. They were also quick to communicate about what was going on, made rides free in that area, and reimburse those affected. Alas, the damage was done. This situation left a black mark on their reputation and serves as a warning to marketing managers that any algorithm they implement needs to be constantly monitored and have the possibility to be overridden built in.

The Paradox of Automation

The purpose of automation is to replace the role of humans, aiming to make tasks faster and more accurate and leaving people free to do more complex work. The downside to this is that then people don’t have experience with those simpler tasks and don’t have the opportunity to gradually build up their expertise and skills. In marketing, this could mean that those in marketing, from customer service agents to market research analysts, miss the opportunity to hone their skills on simpler and more repetitive tasks that allow them to better understand customers and their needs, and are left dealing with only the most complicated and unique cases. It remains to be seen what implications this would have for the quality of service and work.

The next frontier of AI and marketing: transferring and creating knowledge

What sets AI apart from traditional statistics is its ability to execute higher-order learning, like uncovering relationships between indicators to predict the likelihood that an Internet user will click on an ad, and to do so autonomously. Being able to create knowledge like this is a huge advantage of AI. However, the transfer of knowledge from the AI model to the expert and vice versa is a major weakness of AI. Since marketing deals with human behavior, this requires a lot of common sense, which, as we now know, is not the forte of AI models. Since this kind of knowledge is often more implicit, dealing with social codes and norms, it is also harder to program into an AI model. The machine will also be able to pick up on links that it needs to transfer back to the human expert, especially so that the experts can identify flaws in the system and understand how it is operating. An AI system that is able to create and transfer knowledge back to the human expert is thus the Holy Grail of AI technology.

More ESSEC Knowledge articles here

[ESSEC KNOWLEDGE] by Jeroen Rombouts - Professor at ESSEC Business School. He holds the Accenture Strategic Business Analytics chair.

Since the start of the COVID-19 pandemic, companies have had to accelerate their digital transformation. This implies increased investments, so substantial that they require C-level support. The stakes are high for organizations. From accelerating sales to optimizing operational processes, digital impacts the value chain in every aspect. If the digital revolution generates an inevitable modernization of companies and a hope of value generation, it also provokes a major challenge for organizations: Data. Data from transactions, customers, products, etc. invades the daily operations of organizations, constituting a potentially valuable asset, but above all an important challenge in terms of governance and management. Organizations must increase the understanding of these data as part of their transformation. In the very short term and in an uncertain time, data becomes more crucial than ever to identify the levers of performance of companies. Optimizing costs, increasing business revenues, and driving process efficiency are all initiatives based on the availability of relevant data. As the decision cycles accelerate, many decision-makers will no longer be able to drive their businesses with approximate and often inaccurate data. Having good data - and just in time - has become a pressing necessity. But this prospect seems attainable only if the data heritage is better mastered. This is precisely the purpose of the “Data Footprint” method designed by Kearney and Essec. Evaluating the data footprint now constitutes an essential approach to secure investments and increase control over data assets. The Data Footprint approach introduces a virtuous practice that aims to understand the data heritage, risks, challenges and limits linked to data within organizations. The Data Footprint is an evaluation process based on a 360° analysis of the data required as part of a company initiative steered by the entity in charge of Data Governance. The aim of the Data Footprint is to assess the data assets to establish a risk assessment score. Based on multiple dimensions of analysis such as data quality or security, our method allows a quantified assessment of the data heritage in an organization. Today, the data heritage is still poorly controlled and exploited in many companies.

What is the quality level of critical data sets in the organization (e.g customers/ suppliers’ data)? What is the level of risk associated? What is the degree of control and ownership of data in the organization? These questions are often asked by decision makers without concrete answers based on a structured assessment. The complexity of information systems combined with the lack of governance make the data equation often complex and costly. The Data Footprint allows companies to get a tangible data assessment across multiple dimensions in order to establish a risk score. The purpose of such a measure is to be able to accurately assess areas of weakness and to monitor data heritage improvements. The approach also allows internal and external benchmarks based on a standardized analysis grid. The strategy for implementing a Data Footprint should be progressive while focusing on the critical data sets in the context of companies’ major programs, projects or business transformation initiatives. The approach should involve several collaborators, at least representatives of business lines and IT, who jointly use a score sheet based on the following five dimensions: accessibility and availability, quality, ownership, risks, and identification of the future users. The overall score calculated on these five dimensions can range between 0 and 15, the lower the score the higher the risk related to the enterprise initiative. Consider as an example a company specializing in the distribution of electronic equipment to the general public through its distribution network of more than 2,000 stores. As part of its data strategy, the company decides to launch a priority project that deploys a “Customer-centric” approach in order to increase customer value. The objective is to capture a better understanding of customer preferences in order to meet their expectations. The company anticipates a significant potential risk linked to data (availability, quality, etc.) and decides to launch a Data Footprint approach. The total Data risk score for this company was less than 5 in the evaluation exercise. On the recommendation of the Chief Data Officer in agreement with the rest of the team, the decision to launch the project is postponed pending the implementation of a specific data related action plan. This approach allowed the company to apprehend a major risk related to data on this project. Indeed, a rapid launch of this project without prior assessment would have potentially led to failure with economic consequences (losses estimated at a few hundred thousand euros). The approach also made it possible to initiate collaborative work around the data over the entire duration of this assessment (one month), and thus avoiding internal misunderstandings about the responsibilities of the various stakeholders (Business lines, IT teams, etc.). Finally, a clear action plan could be drawn justifying the investment of technical and human resources to upgrade the information system.

More ESSEC Knowledge articles here

[ESSEC KNOWLEDGE] by Julien Malaurent and Thomas Huber - Associate Professors of Information system at ESSEC Business School.

Triggered by an event in your calendar, your smartphone assistant reminds you to leave for work early. News updates tailored to your morning habits pop up on your phone. The daily decision to click on them or not further refines the parameters of your readership profile. You arrive at work, respond to emails, order raw materials that are low on stock according to the ERP system, and then head home to take a run. Over the course of the day, you have left a trace of personal data all over: the words you spoke in your own home recorded by your smart speaker, the conditions of your commute recorded by your phone, the nature, content, and performance of your daily work carefully tracked by enterprise systems, how far and where your run takes you precisely measured by your fitness wearables. Until you go to bed, at which point you will feed your health app with your sleep patterns, you will further add to this data trace: the contents of private text messages, the items of your shopping basket, Google search queries, Facebook profile visits, your choice of mates and dates. Your private and professional activities performed in the digital world, all recorded, rendered as data, analyzed, processed, and exploited. This may sound like the envisioning of a privacy nightmare fraught with ethical dilemmas – but it is in fact the present. It is the state of the world in the age of what Harvard University’s Shoshanna Zuboff calls ‘surveillance capitalism’. Yet this age has only just begun. The next wave of large-scale, AI-powered systems is already in the making and will bring new ethical dilemmas and challenges because they touch on the very core of societal vs. individual wellbeing, algorithmic certainty vs. human agency, and technological determinism vs. moral responsibility. In this article, we briefly present the current status of AI-powered governance systems, and suggest directions to understand how the organizations and individuals that are developing and using these systems could cope with the ethical dilemmas that they may encounter. The current generation of AI-powered systems such as the services offered by Google or Facebook set out to gather, process, and commodify the totality of information about our every thought, word, and deed expressed in the digital world. This entails data such as social media behavior, purchasing habits, and credit card transactions. The next generation of AI-based systems will go far beyond that. As emerging technologies such as context-aware sensors, facial-recognition cameras, and crowd-based reporting regimes become more ubiquitous, there will be an increasingly complete data record not only of our actions performed in the digital but also in the physical world. This data will be on an unprecedented level of granularity and can be used to not only predict individual behavior but also to shape, produce, modify, manipulate, and control it. Not only the private sector with its power - and resourceful tech giants - will be a key player as a key sponsor of these systems. In fact, many nation states have started to work on largescale, AI-powered governance systems. The Chinese Social Credit system is a prime example of these AI-powered governance systems. These systems come with the promise of optimizing almost all aspects of a populace ranging from energy consumption, to health and sleep. By increasing the transparency of human behaviors and quantifying even the most personal acts, nation states can entice a competitive arms race of social desirability and the will to improve between their citizens.

Two Ethical Dilemmas of Large-Scale AI-Powered Systems.

Despite the many promises of such allencompassing AI-powered systems, they also raise important ethical questions, not only from the perspective of individual citizens and the wider society but also from the perspective of the designers and developers of AI powered products and services as well as the managers overseeing design and development. The first ethical dilemma pertains to the gathering and exploitation of data. On the one hand, the creators of AI-powered systems have strong incentives to gather and exploit ever-larger amounts of data. The reason for this insatiable hunger for data is the technical logic of modern AI systems. These systems become more accurate the more data they are fed with. On the other hand, the more the providers of AI systems use this data to shape and control behaviors, the more they restrain human agency. They replace free will and individuality with algorithmic certainty and guaranteed outcomes (Zuboff, 2019). Hence, an important ethical challenge is to balance the wider societal needs that these systems are designed to serve with the individual need for freedom of expression and free choice. A second ethical challenge is that of the moral responsibility for the behaviors of the AI-based system. If an AI shows some undesired behavior such as discrimination and biased decisionmaking, who is responsible for it? Let’s stop thinking that machines are bad or evil. Algorithms do not distinguish between “good” and “bad” people or good and bad behaviors by themselves. Those systems have been produced by IT professionals, i.e., by designers, developers, and engineers. These professionals have fed them with classifying and clustering rules or at least with the datasets that they deemed appropriate to extract these rules inductively. Yet one cannot easily assign responsibility to the engineers alone either. The lines for moral responsibility are blurred by the fact that due to their intrinsic nature, artificially-intelligent systems are to some extent autonomous; they learn to show behaviors that go beyond their original programming.

Investigating the Management of Ethical Dilemmas in the Age of AI.

ESSEC has positioned itself around three strategic pillars: Enlightening Entrepreneurship, Together, and the Institute for AI & Business. The first aims to invest in our entrepreneurship and innovation ecosystem. The second is the creation of knowledge that will help the current and the future generations of leaders to act not only with business acumen but also with social responsibility (Together). The third is the creation of knowledge that will help leaders navigate through the “storm” of technology-driven change processes such as the digital transformation (Institute for AI & Business). We believe that in order to ensure that the next generation of AI-based systems will not only benefit a few private and public organizations but also strengthen the societal institutions that liberal democracies rely upon, the ethical and technical aspects of management need to become two sides of the same coin. Therefore, we have recently started a research project that takes an in-depth © GettyImages - Metamorworks look at one of the largest AI initiatives in the world. In this research project, we ‘unblackbox’ the complex interplay between instrumental managerial considerations such as technical proficiency or economic efficiency with ethical value orientations. Our goal is to understand how both ethical and instrumental considerations become entangled with modern AIpowered systems that weave together smartphones, stationary computers, CCTV systems, and other sensory devices to form a technological apparatus of unprecedented complexity. Through this large-scale investigation, we strive to uncover which ethical dilemmas the actors involved in the development and improvement of these systems encounter and how they cope with these dilemmas. We moreover expect to considerably advance current understanding of how the technical, the managerial, and the ethical intertwine in real-world AI projects and together drive both intended and unintended consequences. Overall, we are hopeful that this research project will produce actionable managerial levers for designing and maintaining large-scale AI-powered systems that are ethics savvy.

More ESSEC knowledge articles here

[ARTICLE]

By Prasanna Tambe, Peter Cappelli & Valery Yakubovich - Professors at ESSEC Business School & Wharton Business School.

The authors consider the gap between the promise and reality of artificial intelligence in human resource management and suggest how progress might be made. They identify four challenges in using data science techniques for HR tasks:

1) complexity of HR phenomena,
2) constraints imposed by small data sets,
3) accountability questions associated with fairness and other ethical and legal constraints, and
4) possible adverse employee reactions to management decisions via data-based algorithms.

They propose practical responses to these challenges and converge on three overlapping principles - causal reasoning, randomization and experiments, and employee contribution—that could be both economically efficient and socially appropriate for using data science in the management of employees.

[Please find the research paper here]

by Isabelle Solal - Assistant Professor of Management at ESSEC, where she teaches courses on entrepreneurship and organisational behaviour.

We tend to think of women supporting women as a good thing: we’re all in this together, so why not give each other a hand? This has been touted as a strategy for reducing the gender gap in many areas, including entrepreneurship, but the jury is out on whether that support might come at a cost. Isabelle Solal, assistant professor of management at ESSEC, and Kaisa Snellman (INSEAD) investigated the impact of investor gender on the success of female entrepreneurs, finding that in fact, pitches by female-backed female entrepreneurs are viewed as lower quality and the entrepreneurs as less competent, and that female founders who received funding from female venture capitalists were less likely to raise additional funding. The road to hell is paved with good intentions - and relying on female investors to support female entrepreneurs may have unintended consequences.

Women in entrepreneurship

Even in 2021, women are still underrepresented in many traditionally male-dominated fields, entrepreneurship being one such field. Policy-makers, scholars, and the press alike have developed different solutions, ranging from legal ones like affirmative action or gender-blind recruitment and selection processes. One solution that has been put forward is to encourage women who have “made it” to support other women through networking, mentoring, sponsorship, and even financial support. While same-gender support in other fields has been shown to have certain psychological benefits, we need to better understand its effects on audience perceptions, especially in a domain like entrepreneurship where perceptions play a significant role. 

Recent research has examined whether the presence of female investors increases the likelihood that female founders receive funding. This may be because women and men alike tend to support their own gender, or because women are actively aiming to boost other professional women. This gender homophily has been linked to strong trust and better communication in past research.  Despite these potential benefits, there could also be unexpected consequences for the female entrepreneurs who receive help from female investors. Why? Other observers might make biased inferences about the relationship, for example assuming that the female investors offered funding to the female entrepreneurs simply because of their gender and not they deserve it- a perception not applied to male investors funding male entrepreneurs. This can take a toll on how the female entrepreneurs’ competence is perceived, potentially resulting in difficulty accessing resources or lower performance reviews. Dr. Solal and Dr. Snellman explored this question using experimental and field data.

What happens when female investors support female entrepreneurs?

In the first part of the study, the researchers studied field data from early-stage startups in the United States. Since young startups need a significant amount of capital, they will seek out investments from venture capitalists who then take a minority stake in the business. This type of funding is done in rounds, depending on the firm’s development stage and funds required, and so venture capitalists will not provide startups with all necessary capital in the first round due to the high failure rate of new business ventures. Using data from Crunchbase, which has a wealth of information gathered from various sources including the media, VC firms, entrepreneurs, and investors, they examined firms that received a first round of VC funding between January 2010 and April 2018, for a total of 290 female-founded firms out of a sample of 2136. 

They found that female-founded firms who received first-round funding from female venture capitalists were two times less likely to receive additional funding compared to female-founded firms that received funding from men, a phenomenon that was not seen in male-founded firms. 

Building on their initial findings, the researchers recruited 134 MBA students from a prestigious business school and had them watch one of four pitch videos, featuring either a male or female entrepreneur, backed by either a male or female investor. They then rated the pitch’s quality and the entrepreneur’s competence: the results revealed that pitches by female-backed female entrepreneurs were seen as lower quality compared to other conditions, and that these lower ratings were linked to perceptions of the entrepreneur’s competence. This result was only seen for the female investor x female entrepreneur pitches - not, say, the male entrepreneur x male investor pitches. 

 Taken together, the results indicate that female entrepreneurs who receive funding from female investors are likely to be perceived as less competent and to be less successful in raising capital past the seed round. 

What does this mean for female entrepreneurs and female investors?

While we can agree that women supporting women can be beneficial, this study highlights the fact that same-gender support can have negative consequences in some situations. In entrepreneurship, where early funding decisions hinge on how potential investors perceive the founder, female entrepreneurs who received funding from female VCs are seen as less competent since observers are less likely to believe that they received the funding due to merit. These women then go on to raise less funding. 

Policy implications: inclusive investing

Female entrepreneurs may face challenges growing their business if they don’t have a male investor.  Additionally, female investors may be pigeon-holed if they are seen as only investing in female-led businesses. As a result, initiatives encouraging women to support women may negatively impact both parties and put undue pressure on women. ew initiatives should focus on encouraging inclusive investing, where men and women work together to support female entrepreneurs.

Entrepreneurship is a male-dominated field, and while efforts to reduce the gender gender gap by encouraging women to support female entrepreneurs are well-intentioned, they may unintentionally backfire in some instances. This suggests that working toward gender equality in entrepreneurship needs to be a team effort, undertaken by men and women alike. 

Read more articles here

Founded in 2020 by ESSEC Business School, The Institute for AI and Business for Artificial Intelligence, Data and Society helps organizations navigate and better understand the social, economic, cultural, and ethical impacts of AI and data

metalab@essec.edu

Learn more about the Institute for AI and Business

copyright © 2026 Institute for AI and Business

↓
Résumé de la politique de confidentialité

Ce site utilise des cookies afin que nous puissions vous fournir la meilleure expérience utilisateur possible. Les informations sur les cookies sont stockées dans votre navigateur et remplissent des fonctions telles que vous reconnaître lorsque vous revenez sur notre site Web et aider notre équipe à comprendre les sections du site que vous trouvez les plus intéressantes et utiles.