U.S. states are leading innovation in data protection law and regulation. Four states (California, Colorado, Connecticut, and Virginia) have enacted laws that require data protection assessments (DPAs), and three states (Indiana, Tennessee, and Montana) have passed legislation requiring DPAs which are awaiting their governors’ signatures. These DPAs consider the benefits to a broad range of stakeholders. The full range of potential adverse processing impacts to consumers, and the mitigation necessary to offset those potential adverse impacts. This is where the innovation lies. Colorado, additionally, has gone a step further and adopted rules which specify the content of the DPAs. Colorado’s new privacy rules go into effect in July, and they are a game changer. Part 8 of the Colorado Rules is entitled “Data Protection Assessments.” Rule 8.02 is entitled Scope, and it states: A data protection assessment shall be a genuine, thoughtful analysis of each Personal Data Processing activity that presents a heightened risk of harm to a Consumer … that : 1) identifies and describes the risks to the rights of a consumers associated with the processing; 2) documents measures considered and taken to address and offset those risks, … 3) contemplates the benefits of the Processing; and 4) demonstrates that the benefits of the Processing outweigh the risks offset by safeguards in place. Notably, the assessment must reach beyond just the individual consumer. Section A.5. of Rule 8.04, which specifies the DPA content, states: The core purposes of the Processing activity, as well as other benefits of the Processing that may flow, directly and indirectly to the Controller, Consumer, other expected stakeholders, and the public. Section A.5 requires the DPA to look not only at the benefits associated with the processing to the controller and the consumer, but other stakeholders as well. To do that the DPA will have to actually catalog who those stakeholders might be. Section A.6 then lists the adverse consequences that the DPA needs to assess against (contained in the table below): Colorado examples of risks to the rights of consumers that may considered in a DPA Constitutional harms, such as speech harms or associational harms Intellectual privacy harms, such as creation of negative inferences about an individual based on what an individual reads, learns, or debates Data security harms, such as unauthorized access or adversarial use Discrimination harms, such as a violation of federal antidiscrimination laws or antidiscrimination laws of any state or political subdivision thereof, or unlawful disparate impact Unfair, unconscionable, or deceptive treatment A negative outcome or decision with respect to an individual’s eligibility for a right, privilege, or benefit related to financial or lending services, housing, insurance, education enrollment or opportunity, criminal justice, employment opportunities, health-care services, or access to essential goods or services Financial injury or economic harm Physical injury, harassment, or threat to an individual or property Privacy harms, such as physical or other intrusion upon solitude or seclusion or the private affairs or concerns of Consumers, stigmatization or reputational injury Psychological harm, including anxiety, embarrassment, fear, and other mental trauma; or Other detrimental or negative consequences that affect an individual’s private life, private affairs, private family matters or similar concerns, including actions and communications within an individual’s home or similar physical, online or digital location, where an individual has a reasonable expectation that Personal Data or other data will not be collected, observed, or used. Section A.7 covers mitigation measures and Section A.8 requires a balancing of benefits against the risks described in Section A.6. and the measure used to reduce those risks. The IAF team has looked at requirements in Europe and other jurisdictions, and none contain the breadth of parties to be considered and a description of what significant risk might entail. On the one hand, there are few companies that have the current capacity for these assessments. On the other hand, Part 8 of the Colorado Rules begins to recognize that processing is not just about the consumer, as a data subject, and the controller. It recognizes that complex processing requires an assessment that looks horizontally, both through the organization and externally, to find the appropriate multi-factor balancing. This type of balancing will be required to bridge the differences between legacy privacy governance systems, and fair advanced processing including machine learning and artificial intelligence (AI). As for innovation, the Colorado rules will be studied and considered by other jurisdictions. The IAF team believes that similar rules on assessments likely will be adopted in California and will cascade from there. The IAF has initiated a new project called the “Colorado Project” that will develop an assessment template based on Colorado Rule 8 and expected future regulations in California. The Colorado Project will include a multi-stakeholder dialog to be held most likely in Colorado. The IAF June retreat in San Francisco also will include a discussion on the impact of these DPAs on the way fair AI is balanced. There Is Privacy Law Innovation in the United States April 30, 2023 Lynn A. Goldstein Articles and News Publications Media

Bermuda Privacy Commissioner Accountability Report March 2020 Home / Publications / Download PDF

IAF Comments to the California Privacy Protection Agency Stakeholder Sessions May 2022 Home / Publications / Download PDF

Vice President, Chief Privacy & Data Ethics Officer at Cognizant Stephanie Higgins Vice President, Chief Privacy & Data Ethics Officer at Cognizant Stephanie is a seasoned global privacy professional with over twenty years of experience managing complex privacy laws and regulations for multinational businesses. Her focus is on privacy strategy including developing programs, policies, and processes necessary to ensure compliance and responsible use of data. She joined Cognizant in January 2019 as Chief Privacy and Data Ethics Officer and leads a global team focused on devising and implementing a global approach to personal information handling. Prior to her current role, she spent 18 years with Deloitte, most recently leading their Global Privacy Office and advising on data protection requirements impacting their global organization. Previously as a regulatory consultant, she specialized in data protection and technology assurance and advisory services supporting multinationals in a range of sectors. Stephanie Higgins Vice President, Chief Privacy & Data Ethics Officer at Cognizant Stephanie is a seasoned global privacy professional with over twenty years of experience managing complex privacy laws and regulations for multinational businesses. Her focus is on privacy strategy including developing programs, policies, and processes necessary to ensure compliance and responsible use of data. She joined Cognizant in January 2019 as Chief Privacy and Data Ethics Officer and leads a global team focused on devising and implementing a global approach to personal information handling. Prior to her current role, she spent 18 years with Deloitte, most recently leading their Global Privacy Office and advising on data protection requirements impacting their global organization. Previously as a regulatory consultant, she specialized in data protection and technology assurance and advisory services supporting multinationals in a range of sectors.

IAF Comments to UK ICO AI Consultation February 2024 Home / Publications / Download PDF

Dun & Bradstreet Dun & Bradstreet

A comprehensive, preemptive, federal privacy law that creates a single set of rules for the United States is a once in a generation effort that will have a lasting impact for decades and is long overdue. The drafters of such a bill, the American Data Privacy and Protection Act (ADPPA), should not force companies and courts down the road to guess what the text means. Given the very limited and narrow rulemaking authority granted to the Federal Trade Commission in the draft ADPPA, it’s even more incumbent on Congress to get the text right so that, on a mechanical level at least, the draft ADPPA’s directives can be followed and enforced. A tremendous amount of work went into drafting the ADPPA. It’s an impressive bipartisan effort on an important and timely issue. The most recent version of the draft ADPPA, however, is difficult to interpret and if enacted into law would be challenging to implement and enforce. This comment is not a criticism of substantive decisions or policy compromises but rather an observation that in many places the text of the draft ADPPA deviates from basic standards for sound legislative drafting, producing an incoherent framework. The draft ADPPA is riddled with vague and ambiguous definitions, undefined terms, inconsistent and imprecise use of different words for the same or similar ideas, and overused, vague modifiers (reasonably, serious, significant, substantial, etc.). All this ambiguity and uncertainty will cause endless legal difficulties, make compliance a guessing game, add to the administrative burden on the Federal Trade Commission and state agencies, hinder enforcement, and undermine the important new rights granted to consumers. I do not want the landmark ADPPA to be bogged down in courts for years as judges attempt to divine the intent of Congress. My goal is not to slam the draft ADPPA, fuel opposition, or derail the legislative effort. To the contrary, I want Members of Congress and stakeholders to redouble their efforts to clarify the language. The good news is that there’s plenty of time if Members of Congress and stakeholders roll up their sleeves and take out their pens. Drafting federal legislation is an arduous task, more difficult than most people appreciate. Guidelines, standards, and conventions for legislative drafting help achieve consistency from statute to statute, making federal laws, at least in theory, easier to read, understand, and follow. These best practices start with the notion that federal laws be “written in plain English for real people.” Although the enacted ADPPA primarily will be read by lawyers and lobbyists, not real people, it still needs significant work if the framework is going to work. Now is the time to complete a line-by-line, word-by-word review of each provision so that the legislative language—the black and white text on the page—is as clear as possible and does what people believe it is intended to do. My article “Cleanup in Aisle ADPPA” has greater detail on making the draft ADPPA clearer. CLEANUP IN AISLE ADPPA March 19, 2023 Marc Groman Articles and News Publications Media

Cognizant Cognizant

IAF Comments to Brazilian LGPD International Transfer Requirements November 2022 Home / Publications / Download PDF

Chief Privacy Officer, Merck & Co., Inc. (USA) Chris Foreman Chief Privacy Officer, Merck & Co., Inc. (USA) Chris is the Chief Privacy Officer at Merck. Heading the Global Privacy Office, an integral part of the Ethics & Compliance Organization, he leads a global team of privacy professionals that oversees the governance and functioning of the Global Privacy Program. Through its standards, specifications, external certifications and guidance, the Global Privacy Office supports a network of 250+ Privacy Stewards embedded within the various operating divisions and global support functions of the Company, and ensures accountability by the business for its activities. Chris spent his first 20 years with the Company in the Office of the General Counsel, thereafter joining the Global Privacy Office in September 2018. He has been Chief Privacy Officer since August 2023. He advocates the Company’s interests externally in several fora, including EFPIA’s Data Governance Working Group, IAF and dplegal. Chris has spoken widely at privacy and legal conferences, webinars and roundtables on topics including the complexities of international data transfers between the Europe and the United States and the European Health Data Space. Before joining the Company, Chris worked as a corporate attorney at two private law firms. He earned his B.A (Government), J.D. and LL.M. from the University of Texas, University of Georgia, and Vrije Universiteit Brussel, respectively. Chris Foreman Chief Privacy Officer, Merck & Co., Inc. (USA) Chris is the Chief Privacy Officer at Merck. Heading the Global Privacy Office, an integral part of the Ethics & Compliance Organization, he leads a global team of privacy professionals that oversees the governance and functioning of the Global Privacy Program. Through its standards, specifications, external certifications and guidance, the Global Privacy Office supports a network of 250+ Privacy Stewards embedded within the various operating divisions and global support functions of the Company, and ensures accountability by the business for its activities. Chris spent his first 20 years with the Company in the Office of the General Counsel, thereafter joining the Global Privacy Office in September 2018. He has been Chief Privacy Officer since August 2023. He advocates the Company’s interests externally in several fora, including EFPIA’s Data Governance Working Group, IAF and dplegal. Chris has spoken widely at privacy and legal conferences, webinars and roundtables on topics including the complexities of international data transfers between the Europe and the United States and the European Health Data Space. Before joining the Company, Chris worked as a corporate attorney at two private law firms. He earned his B.A (Government), J.D. and LL.M. from the University of Texas, University of Georgia, and Vrije Universiteit Brussel, respectively.

Assessments to an AI World: Legitimate Interest Assessment November 2024 Home / Publications / Download PDF

HP Inc. HP Inc.

IAF Comments on Quebec Bill 64- IAF Public- French September 2020 Home / Publications / Download PDF

IAF Comments to the EU Proposed AI Regulation July 2021 Home / Publications / Download PDF

Origins of Accountability: Big Data and Analytics: Seeking Foundations for Effective Privacy Guidance February 2013 Home / Publications / Download PDF

Origins of Accountability: Accountability Phase III – Madrid Project November 2011 Home / Publications / Download PDF

Sun Life Sun Life

Big Data Ethics Initiative: Assessment Framework (Part B) July 2015 Home / Publications / Download PDF

1 CJEU Case in SCHUFA Has Far Reaching Implications Beyond Credit Scoring Martin Abrams, Emeritus Lynn Goldstein, Senior Strategist The European Court of Justice opinion, SCHUFA, that credit scoring constitutes automated decision-making under GDPR Article 22(1) has broader implications beyond credit-scoring. The ruling by the court “to fill a legal gap” implies that the risk scores produced by businesses like fraud detection and identity verification are automated decisions. It suggests controllers will need to obtain consent before calculating creditworthiness or other types of algorithm-based scoring that are used in a wide variety of business processes. The court’s opinion is inconsistent with modern data analytics and well-established credit scoring practices and may be at odds with the evolving role analytic driven decision-making plays in many aspects of life. These analytic processes reflect the concepts “thinking and acting with data.” Thinking with data is the robust use of data to create new insights; use of those insights to affect individuals is acting with data. Although the score in SCHUFA related to a particular individual, until that score was used by a lender – acting with data – that score itself had no impact on an individual. GDPR Article 22 only concerns acting with data. The CJEU overlooks the distinction between thinking and acting with data in order to reach a broad interpretation of the term “decision” in GDPR Article 22(1). Big data were barely understood, and complex analytics were in their infancy, when the GDPR was adopted in 2016. The GDPR is intended to be technology neutral in many respects, but it has some gaps when it comes to regulating advanced analytics. Based on information contained in the order for reference, the court in SCHUFA determines that, in order to fill a legal gap – the data subject cannot obtain access to meaningful information about the logic involved in the score established by credit information agencies from the financial institution the data subject applied for a loan from and the credit information agency is not obliged to provide that information – that score is an automated decision for the purposes of GDPR Article 22(1). In our view, no such gap exists in the GDPR, but even if it did exist, the court should not have presumed what the relationship between the credit information agency and the financial institution is. In doing so, the CJEU reaches an incorrect decision. The GDPR does address how to obtain access to the information at issue here. Usually, controllers and processors enter into agreements which require the processor to assist the controller in responding to such access requests. So, data subjects can obtain access to meaningful information about the logic involved in automated decision-making from the controller, the bank. The issue in the case is what is the relevant decision? The act by which a bank agrees or refuses to grant credit to the applicant? The act by which SCHUFA derives the score from a profiling procedure? The court recognizes that the answer to this question 2 depends on the facts in each case. The problem with the opinion is that the court goes on to make a series of incorrect presumptions about how credit scores are applied to conclude that the credit score is the decision. Ultimately, because of the fact driven nature of the inquiry, the court’s decision may not matter in the financial services industry. However, the broad holding that the court reasoned it should reach because of the absence of a legal definition of the term “decision” in the GDPR means that there are many broader implications for other industries and sectors. For example, scoring is used in retail transactions to identify fraudulent transactions. “Machine learning scores transactions in real time by analyzing factors such as device information, IP address, and location in order to identify potential fraud in ecommerce transactions. If a customer usually pays with a credit card but suddenly switches to a different payment method, it may indicate that their account has been compromised and a real-time notification is sent.” Detecting Retail Fraud Another example is in healthcare. We all are familiar with the scores we receive when we get our blood test results. Are those decisions? The number determines whether a result is diabetes or not. If the doctor solely relies on the score, is the blood test result an automated decision? In the SCHUFA case, if the court’s determination that there is a gap in the GDPR because the data subject cannot obtain access to meaningful information about the logic involved in automated decision-making from the bank because the credit bureau, not the bank, has it, then the court just should have interpreted the law rather than made new law. This judicial activism in unwarranted particularly when the EU AI Act which governs credit scoring will be coming into effect soon. While banks and credit information agencies may be able to get around the holding in SCHUFA because the facts are different, the court’s ruling has implications for other businesses providing AI or other analytical scoring. ANALYSIS OF THE CASE SCHUFA Holding AG is a German credit information agency that provides its clients, financial institutions, with information on the credit worthiness of individuals. SCHUFA provided a financial institution with a score for OQ which served as the basis for the refusal to grant the credit for which OQ applied. OQ then requested SCHUFA to erase the entry concerning her and to give her access to her data, but SCHUFA merely informed her of the score and, in broad outline, of the principles underlying the calculation method for the score, without informing her of the specific data included in that calculation or of the relevance accorded to them in that context, asserting that the calculation method is a trade secret. OQ brought a case against SCHUFA. The court stayed the case and referred to the CJEU for a preliminary ruling on the question of whether GDPR Article 22(1) is to be interpreted as meaning that the automated establishment of a score by the credit 3 information agency concerning the ability of a data subject to service a loan constitutes a decision within the purview of GDPR Article 22(1). CJECU HOLDS CREDIT SCORE IS AN AUTOMATED DECISION In holding that the creation of the score, itself, was an automated decision, the CJEU broadly interprets the term “decision. In determining what is the relevant “decision,” the CJEU observes there is, on the one hand, the act by which a bank agreed or refused to grant credit to the applicant, and on the other hand, the score derived from a profiling procedure conducted by SCHUFA. The CJEU was unable to answer this question because the answer depends on the way in which the decision-making process is structured in each particular case. The CJEU states that this process typically includes several phases: profiling, establishment of the score, and the actual decision on the grant of credit. The CJEU speculates that although a financial institution can take on this process, there is nothing to prevent it from, by contract, assigning certain tasks, such as profiling and scoring, to a credit information agency. The CJEU then incorrectly speculates that the decision-making process could be conceived in such a way that the scoring by the credit information agency predetermines the decision by the financial institution to grant or refuse to grant credit, Thus, if the scoring were carried out without any human intervention that could verify its result and the fairness of the decision with respect of the credit applicant, the CJEU thinks it logical for the scoring itself to constitute the “decision” under GDPR Article 22(1). The CJEU then determines the information contained in the order for reference suggested that the score established by a credit information agency and transmitted to a financial institution generally tends to predetermine the financial institution’s decision to grant or refuse to grant credit to the data subject. Even though the CJEU acknowledges that the facts need to be assessed in each individual case, the CJEU concludes that the score itself is a “decision” within the meaning of GDPR Article 22(1). A GAP IN LEGAL PROTECTION? The CJEU states that it is reasonable to draw this conclusion because a strict reading of GDPR Article 22(1) would give rise to a gap in legal protection. On the one hand, SCHUFA would not be required to provide information to the data subject under GDPR Article 15(1)(h) since it would not be the one making an “automated decision” within the meaning of GDPR Articles 15(1)(h) and 22(1). On the other hand, the financial institution to whom the score is communicated cannot provide information under these Articles because it does not have it and would be unable to review the evaluation of the creditworthiness of the credit applicant if the decision is contested. To avoid this perceived gap, the CJEU proposes an interpretation of GDPR Article 22(1) which it thinks considers the real impact of scoring on the data subject. The CJEU thinks this approach logical as the credit information agency, should, in general, be the only entity capable of responding to requests from the data subject based on the rights guaranteed by GDPR Articles 16 (right to rectification) and 17 (right to erasure), The CJEU wrongly observes that the financial institution generally is not involved in either 4 collecting those data or profiling where those tasks are “assigned” to the credit information agency. There is no gap in the GDPR. The CJEU says that SCHUFA is the only entity capable of responding, but not obligated to respond, to data subject requests under GDPR Articles 15 – 17, and that the only way to solve this gap is to conclude that a score is a decision under GDPR 22(1). The CJEU is incorrect. The CJEU makes incorrect assumptions about the credit information agency – financial institution relationship (the CJEU does not refer to any information contained in the order for reference about the relationship between the credit information agency and the financial institution). This relationship is fact based and must be determined in every case, but generally the financial institution is the controller, and the credit information agency is the processor. When there is a controller-processor relationship, under GDPR Article 28(3), the controller and the processor must enter into a contract that governs the processing the processor does for the controller. Under Article 28(3)(3), the contract must provide that the processor assist the controller in fulfilling “the controller’s obligations to respond to requests for exercising the data subject’s rights laid down in Chapter III.” Articles 15 – 17 are in Chapter III. Therefore, even though the financial institution does not have information about the score, when it receives a request from its customer, the contract it has with the credit information agency requires the credit information agency to assist financial institution in: • Reviewing the evaluation of the creditworthiness of the credit applicant if the decision is contested, and • Responding to requests based on the right of access to data upon which the decision was based, the right to rectification where personal data to carry out scoring proved to be inaccurate, and the right to erasure where those data have been unlawfully processed. Thus, there is no “legal protection gap;” the controller can provide the information when the processor has it because the contract between the controller and the processor requires the processor to assist the controller in providing information in response to requests in Chapter III. CREDIT SCORING AND THE AI ACT The result in SCHUFA is inconsistent with modern data analytics and well-established credit scoring practices. Both of these processes reflect the concepts “thinking and acting with data.” Thinking with data is the robust use of data to create new insights; use of those insights to affect individuals is acting with data. Part of thinking with data is determining the likelihood of an event happening. In developing the scoring mechanism, SCHUFA was a controller but not using data to make a decision on an individual. Credit scores are not stored by bureaus. They are derived at the time of the request. When SCHUFA determined the credit score at issue here, one could argue about whether it was thinking or acting with data. What is not debatable is that in certain factual situations, the credit information agency is acting as a processor for the bank. Although the score 5 related to a particular individual, until that score was used by a lender – acting with data – that score itself had no impact on an individual. GDPR Article 22 only concerns acting with data. The CJEU overlooks the distinction between thinking and acting with data in order to reach a broad interpretation of the term “decision” in GDPR Article 22(1). There is no gap in legal protection; however, if there were, let the new EU AI Act cover it. This broad reading of the term “decision” by the CJEU is unnecessary.. Under the new EU AI Act, high-risk AI systems are those that pose significant risk to fundamental rights, such as those used for credit scoring. High-risk AI systems must comply with strict rules on data quality, transparency, human oversight, accuracy, robustness, and security. Rather than shoehorn the scoring practices at issue in SCHUFA under the GDPR, let the new AI Act come into effect and let the practices at issue in SCHUFA be governed by it. Not every issue involving personal data must go through the GDPR. The GDPR does address how to get the information at issue here, but even if it did not, then the new AI Act addresses how to get it. CREDIT SCORING Credit scoring has existed for a longer period of time in the U.S. than in the EU. Some learnings in the U.S. have relevance. When David Medine was director of financial practices at the FTC, he said he preferred decisions based on credit scores to those made by lending officers with possible prejudices, Time has shown that scoring expanded credit further and deeper into populations. The inconsistent data pertaining to populations baked prejudice into the process. It was and is a data issue. A credit score is a tool to make better decisions. Not perfect decisions, but better decisions. Credit scores are based on probability. The logic could be explained by saying that if there were a hundred consumers whose data looked like you, x number would go bad over a determined period of time. Bad could be a credit default or a significant delinquency. Explaining the logic in those terms is a doable task by the model developer. The concept of scoring for significant decisions has been more sensitive in the EU than the U.S. That is why making the logic transparent is important. However, defining the creation of the science behind the score as decisioning has ramifications. GDPR Articles 9 and 89 come into play and impede conducting the science. Scoring has been sensitive in Europe for over 25 years for several reasons. First, the protection of human dignity – preventing the data subject from being subject to a decision based solely on automated processing. Second, the data in Europe was a negative, not full, file. There is no gap in the GDPR. Going beyond the information contained in the order for reference and making incorrect assumptions about the credit information agency - financial institution relationship led the CJEU to broadly interpret the term “decision” in GDPR Article 22(1) in order to address a nonexistent gap. Even if there were a gap, it is not unusual for gaps to exist in legislation; there is nothing wrong in not having anticipated every possible use of technology when the GDPR was drafted, especially 6 when new legislation, the AI Act, is awaiting final passage that will address this new technology. CJEU Case in SCHUFA Credit Scoring- Policy Analysis December 2023 Home / Publications / Download PDF

The following blog was taken directly from the IAF comments filed in response to the California Privacy Protection Agency request for comments on assessments and automated decision – making . The February 10, 2023, Invitation for Preliminary Comments asks a series of questions related to automated decision-making and profiling. The IAF is not responding to the specific questions but instead setting forth some basics for the discussion. The fact is that automated decision-making is baked into how things work on an everyday basis. For example, the CPPA uses automated decision-making on requests from browsers to access the CPPA’s servers on a daily basis. These decisions have the effect of limiting who can browse the CPPA’s website and file complaints. This is good because the alternative would be constant security breaches. However, the issues related to profiling and automated decision-making predate when consumer browsers made the Internet a consumer medium. Martin Abrams, former President and current Chief Policy Innovation Officer of the IAF, was the President of the Centre for Information Policy Leadership (CIPL), the Vice President, Information Policy, Experian, Director Consumer Policy, TRW Information Systems and Services and the Community Affairs Officer of the Cleveland Federal Reserve Bank. His background gives him the perspective to provide the following comments. The consumer Internet accelerated an observational age that in turn accelerated the use of data for probabilistics pertaining to how people behave. The first broad-based probabilistic use of consumer data was probably the Fair Isaac credit risk score in 1989. It was quickly adopted by the consumer lending industry as an aid to better decisioning than was possible with the subjectivity of decisions made purely by lending officers. Soon that aid to people evolved into automated credit decisions. The U.S. Department of Justice (DOJ) investigated whether those decisions had the effect of making decisions on grounds that violated the Equal Credit Opportunity Act (ECOA). Since the data for credit risk scores came directly from credit bureaus, the FCRA required that the use of scores must be disclosed along with the factors that led to the denial. So, from the very beginning, the use of profiling and automated decision-making for substantive decisions were covered by a fair processing law, the FCRA. In Europe, there was no uniformity in the data available for consumer credit decision-making. As Europe evolved towards the creation of the 1995 EU Privacy Directive, there were debates on whether it was unseemly for decisions on people to be made solely by a machine. Those concepts on what is seemly or not influenced the drafting of Article 22 of the GDPR. So, there are cultural differences between the way that Europe sees these issues and the way they are seen in the United States. The fact is that the relationship between profiling, the use of probabilistics against broad data sets, and automated decision-making is muddled still under Article 22 of the GDPR. The 21st century saw the rise of analytic skills that allowed for the use of unstructured data into advanced analytic processes. Legacy statistics tested causality, while the growth of big data switched the dominant theme to correlation. This change naturally raised questions about the accuracy of the correlations, whether they were appropriate to apply, and whether they were influenced by the bias built into available data sets. This development has informed the debate about algorithmic fairness. These concerns have accelerated with the growing use of AI, which is the next stage of advanced analytics in our observational world. So, in thinking about the questions the CPPA is asking, some pragmatic truths need to be addressed: Profiling is probabilistics built with consumer data. Building choice into the data that feeds the probabilistics has the unintended consequences of skewing the accuracy of predictive values. Choice worked when the relationship was one on one. Most relationships are no longer one on one. Ours is an observational world where there are not many one-on-one relationships. Choice no longer fits and indeed harms the process in an observational world. Automated decision-making is built into how many modern processes work, including the functioning of the CPPA’s cybersecurity processes. Many automated decision-making processes are subject already to laws such as the FCRA, ECOA, and Fair Housing Act (FHA). The FCRA, ECOA, and FHA wrestled with these issues already and decided that the benefits of the automated decision-making outweighed the risks. Those Acts have methods for determining whether the automated decision-making is biased or not (after the fact testing), and those methods are just as applicable today as they were when they were implemented. Much of the emotions that pertain to automated decision-making are related directly to whether one thinks it is fairer for a person to make a decision or whether a well-governed algorithm, in the end, would be fairer. As mentioned above, the DOJ in the context of the ECOA decided that a well-governed algorithm was better. The IAF staff believes this is where the discussion should begin. Automated Decision-Making and Profiling are Not New Issues March 28, 2023 Martin Abrams Articles and News Publications Media