Ethical values elicitation is a structured process for identifying the values, concerns, and expectations that should shape a technology system before it is designed, deployed, or assessed. In the AI context, it helps organizations move beyond broad principles such as fairness, accountability, privacy, and transparency and ask a more precise question: what do those values require in this specific system, for these stakeholders, in this operating environment? That question sits at the center of IEEE CertifAIEd™ AI Ethics and the value-based engineering approach reflected in IEEE 7000-2021, which describes processes for addressing ethical concerns during system design.
For businesses, ethical values elicitation is not simply an ethics workshop or a general stakeholder listening exercise. It is a disciplined way to surface ethical expectations, prioritize competing concerns, and translate those concerns into requirements that can be considered by product, engineering, legal, compliance, risk, and executive teams.
Why Ethical Values Need to Be Elicited, Not Assumed
Most organizations already use some form of requirements gathering. Product teams interview users, engineers define performance targets, compliance teams interpret legal obligations, and executives set business objectives. Ethical values elicitation adds a different layer. It asks what a system should respect, protect, or avoid from the perspective of the people and institutions affected by it.
That distinction matters because ethical terms often look simple until they have to be implemented. “Transparency,” for example, can mean different things to different stakeholders. A regulator may expect documentation and auditability. A customer may expect a clear explanation of how a decision was reached. A technical team may focus on model interpretability or traceable data lineage. Each version is related to transparency, but each points toward a different design requirement.
IEEE 7000™ addresses this gap by establishing processes through which organizations can include ethical values during concept exploration and development. The standard specifically supports transparent communication among management, engineering teams, and selected stakeholders for ethical values elicitation and prioritization, with traceability through the operational concept, value propositions, and system design. That traceability is what distinguishes values elicitation from a less formal discussion about responsible innovation.
How Ethical Values Elicitation Works in Practice
A values elicitation process usually begins by defining the system context. The organization clarifies what the AI system is intended to do, who will use it, who may be affected by it, what decisions it may influence, and what organizational or regulatory constraints are already known. This stage is important because ethical risks are rarely abstract. They emerge from a system’s specific purpose, setting, users, data, and consequences.
The next step is stakeholder identification. Stakeholders may include end users, customers, employees, impacted communities, product owners, developers, compliance teams, procurement teams, regulators, and external partners. In some systems, the most important stakeholders are not the direct users. A business may deploy an AI tool for internal decision support, but the people most affected could be job applicants, patients, borrowers, students, or members of the public.
Once stakeholders are identified, the organization uses structured interviews, workshops, scenario analysis, or facilitated discussion to elicit values and concerns. The goal is not to collect general opinions about AI. It is to identify ethically significant expectations that should influence the system’s design and governance. Stakeholders might raise concerns about contestability, privacy, bias, accessibility, explainability, human oversight, data retention, or the appropriate boundaries of automation.
The most important step is translation. Elicited values must be turned into value qualities, ethical requirements, and system requirements. For example, a stakeholder concern about “being able to challenge an automated decision” may translate into requirements for notice, explanation, human review, response timelines, documentation, and escalation pathways. A concern about privacy may translate into data minimization, retention limits, access controls, or limits on secondary data use. Without that translation, values remain aspirational rather than operational.
Where Businesses Use Values Elicitation
Ethical values elicitation is useful wherever AI systems affect people, rights, access, opportunity, safety, or trust. In financial services, it can help teams clarify how fairness, explainability, and contestability should apply to credit, fraud detection, underwriting, or customer segmentation systems. In healthcare, it can help distinguish what clinicians need from an AI tool from what patients need to understand, consent to, or challenge its use. In employment, it can surface concerns about bias, accessibility, explainability, and the appropriate role of human judgment in hiring or workforce decisions.
The method is also relevant outside heavily regulated sectors. A retailer using AI for personalization may need to consider how customers understand profiling and consent. A manufacturer using predictive systems may need to consider worker safety, accountability, and human override. A software provider embedding AI assistants into enterprise tools may need to examine data exposure, documentation, and user reliance. In each case, the central question is not only whether the system works, but whether it works in a way that reflects the values and expectations of affected stakeholders.
It also supports readiness for emerging AI regulation. The EU AI Act uses a risk-based approach, and the European Commission describes high-risk AI systems as those that can affect safety, health, or fundamental rights in sensitive contexts such as critical infrastructure, education, employment, law enforcement, and migration. While values elicitation is not a substitute for legal compliance, it can help produce the documented reasoning, stakeholder awareness, and design traceability that responsible AI governance increasingly requires.
Managing Conflicts Between Values
One of the most useful aspects of ethical values elicitation is that it makes value conflicts visible early. AI systems often require trade-offs. Greater transparency may create tension with security or intellectual property considerations. Additional data may improve model performance while increasing privacy risk. Automation may improve consistency while reducing opportunities for human judgment or appeal. A system optimized for aggregate accuracy may still raise concerns if its performance varies across affected groups.
Values elicitation does not remove these conflicts. It makes them explicit so they can be evaluated, prioritized, and documented. That is important for business decision-making because AI governance gaps often arise when organizations implicitly prioritize one value, such as efficiency or predictive accuracy, without clearly examining what that choice means for fairness, autonomy, privacy, or accountability.
A mature elicitation process therefore includes prioritization. Stakeholders and decision-makers examine which values are most important in the system’s context, where trade-offs exist, and what safeguards are needed. The outcome should not be a vague statement that all values matter equally. It should be a reasoned record of the organization’s decisions, the rationale behind them, and the requirements or controls that follow from them.
What Businesses Gain From the Process
For business leaders, the value of ethical values elicitation is practical. It helps convert AI ethics from a policy statement into a repeatable governance activity. The process can strengthen requirements definition, improve cross-functional alignment, reduce the likelihood of late-stage redesign, and create clearer evidence for audits, assessments, and external review.
It also helps teams avoid a common governance challenge: separating ethical review from engineering practice. If ethical review happens only after a system has been built, the organization may discover that important concerns are expensive or difficult to address. Values elicitation is most effective when it occurs early enough to influence system concept, data practices, model design, user experience, monitoring, and oversight.
That is why training and assessment pathways matter. The IEEE CertifAIEd™ Practical Assessment Master Class is positioned as a hands-on course for applying the IEEE CertifAIEd process, including scoping projects, creating a Concept of Operations, facilitating an Ethical Values Elicitation, clustering elicited values to criteria suites, and creating the Case for Ethics. For organizations seeking to move responsible AI practices from policy language into assessment-ready work, this kind of structured methodology helps build repeatable capability.
From Responsible AI Language to Responsible AI Evidence
Ethical values elicitation gives businesses a concrete way to answer a central governance question: How did this organization decide what ethical AI means for this particular system? A credible answer requires more than citing principles. It requires evidence that relevant stakeholders were considered, values were identified and prioritized, trade-offs were examined, and system requirements were shaped by that work.
That is where IEEE CertifAIEd™ AI Ethics and IEEE 7000-2021 are especially relevant. Together, they point toward a more disciplined model of responsible AI: one in which ethical considerations are elicited, translated, documented, and assessed as part of the system life cycle. For businesses, that approach can support clearer design decisions, stronger governance practices, and more credible engagement with customers, regulators, and stakeholders.
