Ethics Readiness Levels (ERLs)
- Ethics Readiness Levels (ERLs) are structured frameworks that assess a technology's preparedness for detailed ethical evaluation and intervention.
- They mirror Technology Readiness Levels by aligning ethics—from meaning-oriented to outcomes-oriented—with varying degrees of epistemic and technical maturity.
- ERLs are applied across domains such as AI, quantum technologies, and mathematics education, enabling iterative ethical assessments and multi-stakeholder engagement.
Ethics Readiness Levels (ERLs) are structured, stage-sensitive frameworks for assessing and guiding the incorporation of ethical reflection and intervention across the development, deployment, and education around technologies, mathematics, and artificial intelligence. ERLs are modeled on the established paradigm of Technology Readiness Levels (TRLs), introducing a granular or categorical scale to indicate the preparedness of a given technology, educational approach, or AI system for substantive ethical scrutiny. By calibrating ethical engagement to the epistemic and sociotechnical uncertainties characteristic of different development stages, ERLs align ethical interventions with concrete points of maximum relevance and tractability (Jong, 4 Apr 2025, Adomaitis et al., 10 Dec 2025, Rycroft-Smith et al., 2022).
1. Foundational Definition and TRL Analogy
Technology Readiness Levels (TRLs) are a nine-point ordinal scale used to benchmark technological maturity, ranging from initial observation of basic principles (TRL 1) to full operational deployment (TRL 9). Ethics Readiness is defined as “the extent to which a technology is prepared to undergo detailed ethical scrutiny” (Jong, 4 Apr 2025). ERLs may mirror the granularity of TRLs, as a monotonic mapping , or adopt a coarser level distinction (e.g., four- or six-level schemes).
Within ERLs, a low readiness score designates high uncertainty and limited empirical grounding, with ethical engagement largely conjectural or value-driven. Mid-levels indicate emerging empirical specificity, enabling the application of hybrid or scenario-based ethical approaches. High ERLs correspond to concrete, empirically-demonstrable systems, justifying detailed outcome- and impact-driven analysis (Jong, 4 Apr 2025, Adomaitis et al., 10 Dec 2025).
2. Ethical Modes and Evolution across ERLs
ERL frameworks distinguish between two overarching approaches to technology ethics, shifting in applicability with increasing readiness (Jong, 4 Apr 2025):
- Meaning‐Oriented Ethics: Emphasizes the analysis of values, imaginaries, visions, and the upstream meaning attached to nascent technologies. Appropriate under conditions of epistemic and design uncertainty (ERL 1–3), with methodologies such as Vision Assessment and Hermeneutic Technology Assessment.
- Outcomes‐Oriented Ethics: Targets the articulation and mitigation of specific future consequences, leveraging empirical prototypes or deployed systems. This mode gains methodological traction as technical and usage uncertainties recede (ERL 6–9), enabling structured impact assessment, techno-moral scenario analysis, and regulatory planning.
The following logic formally captures this shift:
This piecewise alignment is pivotal for minimizing both premature speculation and belated regret in technology ethics (Jong, 4 Apr 2025).
3. ERL Formalisms and Implementation: Technology, AI, and Education
ERL frameworks have diverged across sectors to accommodate differences in institutional context and the granularity of readiness.
3.1. Technology ERLs
ERLs in technology map one-to-one to the canonical TRL scale; assignment criteria for each ERL are derived directly from the prototypical TRL definitions—degree of prototype maturity, testing environment, empirical data availability, and deployment status (Jong, 4 Apr 2025).
| ERL | Approx. TRL | Dominant Ethical Concern | Recommended Method |
|---|---|---|---|
| 1 | 1 | Whose vision, meaning attribution | Meaning-oriented |
| 3 | 3 | Underlying value assumptions | Meaning-oriented |
| 5 | 5 | Emerging harms/benefits | Mixed |
| 7 | 7 | Real-world impacts | Outcomes-oriented |
| 9 | 9 | Societal monitoring and regulation | Outcomes-oriented |
3.2. ERLs for Artificial Intelligence
In AI systems, ERLs have been operationalized as a four-level, iterative scale, with each level tied to the “identification–characterization–harmonization–control” cycle and reinforced through a dynamic, indicator-based audit process (Adomaitis et al., 10 Dec 2025):
| ERL | Description |
|---|---|
| 0 | No ethics considerations |
| 1 | Identified ethics/privacy/legal issues |
| 2 | Characterized interactions/trade-offs among issues |
| 3 | Ethics by design: design choices actively harmonize values |
| 4 | Ethics controls: audits, certification, accountability |
Evaluation leverages a tree-structured, block-based questionnaire, with block and indicator selection aligned to technical features, legal regime, and domain specificity (e.g., GDPR, LEA, Robotics). ERLs in this context are dynamically and repeatedly re-evaluated as projects evolve.
3.3. ERLs in Mathematics Education
For mathematics educators, a six-level ERL scale (Levels −1 to 4) adapted from Chiodo & Bursill-Hall targets the spectrum from active resistance to systemic advocacy (Rycroft-Smith et al., 2022):
| Level | Name | Summary Description |
|---|---|---|
| −1 | Obstruction | Active denial or suppression of ethical reflection |
| 0 | Neutrality | Unawareness or disregard for ethical dimensions |
| 1 | Awareness | Recognition of ethical issues but tentative action |
| 2 | Intervention | Initial practical engagement (e.g., lesson design) |
| 3 | Leadership | Formalized, ongoing advocacy in policy and curriculum |
| 4 | Accountability | Systemic critique and reform; calls out harmful practices |
Assignment is based on explicit behavioral and institutional indicators, with no quantitative rubric but extensive qualitative descriptors and illustrative scenarios.
4. Methodologies for ERL Assessment and Scoring
ERL assessment methodologies are context-dependent and vary in formality.
- Technology and AI: Scoring in AI ERLs is implemented mathematically: the cumulative ERL score is initialized at 4.0 and updated by weighted indicator responses (Yes/No) across all relevant blocks; global ERL if , with normalization for sub-domains (“minimum rule”). Weights for each indicator are team-determined and iteratively refined (Adomaitis et al., 10 Dec 2025).
- Education: ERL as applied to mathematics teaching is inherently qualitative; reflection, narrative portfolios, and scenario analysis augment indicator tracking, with progression through levels mapped by task engagement, discourse, and formal policy participation (Rycroft-Smith et al., 2022).
AI evaluation specifically pairs an ethics expert and technical stakeholder, employs tree-structured dynamic questionnaires, and operationalizes high-level values as concrete design checkpoints (e.g., “bias mitigation integrated into training pipeline,” “system supports third-party audit”) (Adomaitis et al., 10 Dec 2025).
5. Empirical Applications and Case Studies
The ERL framework’s utility has been demonstrated in diverse contexts:
- Quantum Technologies: For quantum computing and quantum communication, with low TRL/ERLs, proposed ethical focus is meaning-oriented, critiquing social imaginaries (e.g., “unbreakable encryption”) rather than engaging in outcomes speculation (Jong, 4 Apr 2025). In contrast, quantum sensing at higher ERLs warrants outcomes-oriented analysis regarding surveillance or dual-use impacts—fields that are comparatively neglected by existing scholarship.
- AI Law Enforcement Tools: For a facial sketch generator for policing, initial scoring yielded ERL 0 (“no ethical reflection”), which post-intervention (establishing audit processes, security patching, user training) increased to ERL 3 (“ethics by design”) (Adomaitis et al., 10 Dec 2025). This case underscores the ERL methodology’s capacity to catalyze and track iterative improvements in ethical practice.
- Mathematics Teaching: The six-level ERL taxonomy structures professional development, curriculum audits, and leadership cultivation in mathematics education, serving both as a reflective prompt and as the basis for policy evaluation (Rycroft-Smith et al., 2022).
6. Theoretical Foundations and Practical Implications
ERLs instantiate “translational ethics”—the procedural movement from abstract ethical principles to practice-embedded checks within concrete workflows. In technology and AI, this manifests as the operationalization of normative values (e.g., fairness, transparency, robustness, accountability) at discrete checkpoints (e.g., trade-off analysis, auditability, risk scoring) (Adomaitis et al., 10 Dec 2025). In mathematics education, ERLs are grounded in critical mathematics traditions and challenge the myth of mathematical neutrality, linking individual reflection to institutional critique and systemic change (Rycroft-Smith et al., 2022).
Key implications include:
- Alignment of Ethics and Technical Maturity: Ethical approaches must dynamically calibrate to the epistemic and technological uncertainties characteristic of each stage. Premature outcomes speculation is avoided at low ERLs, while symbolic critique is deprioritized as systems concretize.
- Programmatic Evaluation: ERLs enable recurrent, quantifiable (in technology/AI) or qualitatively structured (in education) auditing of ethical progress.
- Distributed Stakeholding: ERLs formalize the role of multi-actor dialogue—between ethics specialists, engineers, and domain practitioners—in driving iterative improvement and accountability.
7. Best Practices, Limitations, and Evolving Frameworks
Best practices identified across ERL implementations include paring down indicator sets for operational clarity, dedicating resources to domain-specific block creation (e.g., GDPR for AI systems handling data), and implementing paired ethics-technical assessments to circumvent self-assessment biases (Adomaitis et al., 10 Dec 2025). In mathematics education, embedding case-based and reflective exercises anchored to ERL levels in teacher professional development contributes to equity and justice-oriented pedagogical outcomes (Rycroft-Smith et al., 2022).
A plausible implication is that as technological and regulatory contexts evolve, ERLs themselves must be iteratively adapted—both in indicator content and in scoring weights. This flexibility ensures continued relevance and grounding in real-world practice.
Ethics Readiness Levels thus constitute a family of structured, stage-calibrated frameworks for aligning ethical scrutiny with the technical and institutional realities of evolving technologies, AI systems, and pedagogical practice, offering operational guidance to maximize the specificity, tractability, and legitimacy of ethical intervention (Jong, 4 Apr 2025, Adomaitis et al., 10 Dec 2025, Rycroft-Smith et al., 2022).