Ethical Data Management Policy
June 8, 2021: version 0.1
The purpose of this Policy is to outline rules for ethical management of data, including production, processing, and disclosure.1 It is intended for data processors (data analysts and researchers) within KI Design or on its behalf. The Policy applies to all KI Design data, irrespective of source, and irrespective of the recipient of the analytics or research.
For the purposes of this Policy:
- Data is defined as “recorded information used to develop or test human knowledge.”2
- Data analysts and researchers are jointly and individually defined as data processors.
Relevant Ethical Concepts
The following four basic concepts help clearly state the ethical problems — such as the replicability of data — associated with the lifecycle of research data. KI Design must ensure:
- Protection of data subjects’ rights
- Positive balance of impact
- Ethical conduct
in all its research and data analysis projects.
Integrity has two requirements:3
- The quality of being honest and having strong moral principles; and
- The state of being whole and undivided.
Integrity refers both to the character of the data processors as well as to the quality of data. KI Design data processors have an ethical duty to truthfully report their research findings and should never deceive their audience.
The second requirement refers to data integrity: its accuracy and reliability. KI Design data processors must abide by these principles. According to the Committee on Science, Engineering, and Public Policy, “researchers have a fundamental obligation to create and maintain an accurate, accessible, and permanent record of what they have done in sufficient detail for others to check and replicate their work.”4
Protection of data subjects’ rights
Members of the public have a right to information about themselves; they should control whether and how such information is collected, shared, and used. For certain projects or programs, KI Design may seek consent to acquire this right. When research subjects consent to have their personal information collected, there remain requirements with regard to who should have a right to use the research findings, and who should have a right to the potential benefits generated from the research.
What are the positive impacts of the proposed research, and are there any negative impacts? KI Design must identify the various stakeholders and how they are affected, and then maximize the positive impacts while minimizing the negative impacts.
KI Design’s ethical conduct policy requires adhering to the principles of rigor, objectivity, and robustness. In order to uphold ethical conduct in managing research data, it is important to recognize and avoid research practices that are broadly considered “questionable.” KI Design will avoid the following practices with regard to data management:
- Failing to retain significant research data for a reasonable period;
- Maintaining inadequate research records, especially for results that are published or are relied on by others;
- Refusing to give peers reasonable access to unique research materials or data that support published papers;
- Using inappropriate statistical or other methods of measurement to enhance the significance of research findings;
- Failure to adequately test AI tools for bias; or
- Misrepresenting speculations as fact or releasing preliminary research results, especially in the public media, without providing sufficient data to allow peers to judge the validity of the results or to reproduce the experiments (cited in Pascal (2006)).5
Ethical conduct is contractually required of KI Design staff, agents, and third parties; as well, these rules cannot be violated by any tools or technologies used by KI Design.
Evaluating Ethical Conduct
KI Design will utilize the following twelve-step approach6 in analyzing use cases for ethical data management.
- State the nature of the ethical issue you’ve initially spotted
- List the relevant facts.
- Identify stakeholders.
- Clarify the underlying values (including sex, gender, race, ethnicity, age, and other identifiers).
- Consider consequences.
- Identify relevant rights and duties.
- Reflect on which requirements apply.
- Consider relevant relationships.
- Develop a list of potential responses.
- Use moral imagination to consider each option based on the above considerations.
- Choose the best option.
- Consider what could be done in the future to prevent the problem.
Ethical Data Management Requirements
Ethical handling of research data requires not only acting according to ethical principles but also learning about the policies, resources, procedures, and practices that contribute to the collection, analysis, sharing, and preservation of data.
KI Design commits to the following requirements for ethical data management:
- careful planning,
- effective communication,
- proper training, and
- clearly defining and distributing responsibility in the data management or research group.
Guidelines for data management must be included in the research design or data analytics design stage. KI Design analysts or researchers shall outline:
- what types of data are needed,
- how data will be generated and collected,
- what methods will be used for data analysis, and
- how the data will be used to answer research questions.
The guideline must also clearly specify different data processors’ responsibilities in ensuring the quality of data:
- What mechanisms should be utilized to monitor the data quality?
- What channels of communication should be followed if any questions about data management arise?
For each project, an explicit policy for communication must be established during the planning stage. The communication protocol should include stakeholders and their roles in case of a complaint or situation that needs resolution.
It is crucial to provide proper training so that all data processors have a shared understanding of the expectations, standard procedures, and best practices for handling data. Topics will include, but not be limited to:
- best practices for collecting and documenting data,
- proper data analysis, and
- data sharing and license.
Defining and distributing responsibility
Ethical management of research data requires every data processor to clearly understand and meet their responsibility. It is very important that different data processors’ responsibility with regard to data quality should be clearly defined and communicated.
Failure to comply with this policy may result in disciplinary action up to and including termination of employment or association with KI Design, and may also result in legal action being taken by KI Design.