Isaca CDPSE - Certified Data Privacy Solutions Engineer

The primary consideration to protect against privacy violations when utilizing artificial intelligence (AI) driven business decisions is ensuring proper data sets are used to train the models. AI is a technology that enables machines or systems to perform tasks that normally require human intelligence, such as reasoning, learning, decision making, etc. AI relies on large amounts of data to train its models and algorithms to perform these tasks. However, if the data sets used to train the models are inaccurate, incomplete, biased, or outdated, they can result in privacy violations, such as discrimination, profiling, manipulation, or harm to the data subjects. Therefore, an IT privacy practitioner should ensure that the data sets used to train the models are proper, meaning that they are relevant, representative, reliable, and respectful of the data subjects’ rights and interests. References: : CDPSE Review Manual (Digital Version), page 141

Lawfulness and fairness is a data protection principle that states that personal data should be processed in a lawful, fair, and transparent manner in relation to the data subject. This means that personal data should be collected and used for legitimate purposes that are specified and communicated to the data subject, and that respect the rights and interests of the data subject. By posting its customer data protection notice that includes a statement indicating information is collected on how products are used, the content viewed, and the time and duration of online activities, an online business is applying the lawfulness and fairness principle. The online business is informing the customers about the purpose and scope of data collection, and obtaining their consent or legal basis for processing their personal data. References: : CDPSE Review Manual (Digital Version), page 2

Input validation controls are the best way to ensure consumer credit card numbers are accurately captured. Input validation controls are methods that check the format, type, range, and length of the input data before accepting, processing, or storing it. Input validation controls can help prevent errors, fraud, or data loss by rejecting invalid, incomplete, or malicious input. For example, input validation controls can verify that a credit card number follows the Luhn algorithm1, has the correct number of digits2, and matches the card issuer’s prefix3. Input validation controls can also prevent SQL injection attacks4 or cross-site scripting attacks5 that may compromise the security and privacy of the data.

Input reference controls, access controls, and reconciliation controls are also important for data quality and security, but they do not directly ensure the accuracy of consumer credit card numbers. Input reference controls are methods that compare the input data with a predefined list of values or a reference table to ensure consistency and validity. For example, input reference controls can check if a country name or a postal code is valid by looking up a database of valid values. Access controls are methods that restrict who can access, modify, or delete the data based on their roles, permissions, or credentials. For example, access controls can prevent unauthorized users from accessing or tampering with consumer credit card numbers. Reconciliation controls are methods that compare the data from different sources or systems to ensure completeness and accuracy. For example, reconciliation controls can check if the transactions recorded in the accounting system match the transactions processed by the payment gateway.

Data sanitization is a process of permanently erasing or destroying data from a storage device or media to prevent unauthorized access or recovery of the data. Data sanitization methods can include physical destruction, degaussing, overwriting, encryption or cryptographic erasure. The most important consideration when using advanced data sanitization methods to ensure privacy data will be unrecoverable is the type of media on which the data is stored, as different media types may require different methods or techniques to achieve effective sanitization. For example, physical destruction may be suitable for optical disks or tapes, but not for solid state drives (SSDs) or flash memory devices. Degaussing may be effective for magnetic disks or tapes, but not for optical disks or SSDs. Overwriting may work for hard disk drives (HDDs) or SSDs, but not for tapes or optical disks. Encryption or cryptographic erasure may be applicable for any media type, but may require additional security measures to protect the encryption keys or certificates. The other options are not as important as the type of media when using advanced data sanitization methods. Subject matter expertise may be helpful, but not essential, as long as the appropriate method is selected and applied correctly. Regulatory compliance requirements may influence the choice of method, but not necessarily determine it, as different methods may meet different standards or criteria. Location of data may affect the feasibility or cost of applying a method, but not its effectiveness or suitability., p. 93-94 References: : CDPSE Review Manual (Digital Version)

Attribute-based access control (ABAC) is the best approach for limiting the access of regional HR team members to employee data only within their regional office, because it allows for fine-grained and dynamic access control based on attributes of the subject, object, environment, and action. Attributes are characteristics or properties that can be used to describe or identify entities, such as users, resources, locations, roles, or permissions. ABAC uses policies and rules that evaluate the attributes and grant or deny access accordingly. For example, an ABAC policy could state that a user can access an employee record if and only if the user’s role is HR and the user’s region matches the employee’s region. This way, the access control can be tailored to the specific needs and context of the organization, without relying on predefined or fixed access levels.

Ensuring appropriate data classification should be done next after a migration of personal data involving a data source with outdated documentation has been approved by senior management, as it helps to identify the types, locations, and owners of the data, and to apply the appropriate privacy controls and measures based on the data classification level. Data classification also facilitates the data discovery, data minimization, data retention, and data disposal processes15. References: 1 Domain 3, Task 2; 5 Page 9