Open research questions in Ethics and Social Impacts of AI

What the literature leaves open

• Existing research identifies that AI governance readiness varies significantly across organizations and regions, but does not address how middle powers—particularly those with emerging AI sectors in Global Capability Centres—can develop institutional capacity and governance maturity to participate effectively in global AI governance negotiations and standard-setting processes.

  Racing Ahead, Governing Behind: An Institutional Analysis of AI Governance Readiness in Global Capability Centres · 2026 · DOI

• The literature documents systematic divergence in AI governance models across regions (rights-based, market-driven, state-centric, hybrid, developmental) but lacks empirical research on how middle powers can leverage their distinct governance approaches to build coalitions or coordinate positions in global AI governance forums, or how their policy models might inform more inclusive multilateral frameworks.

  Artificial intelligence governance and social policy divergence: a comparative political economy perspective on global AI regulation · 2026 · DOI

• As governments increasingly deploy artificial intelligence (AI) for automated decision-making, little is known about why citizens differ in their attitudes toward AI-driven decision-making in the public sector, especially in welfare governance, where decisions carry profound consequences for marginalized groups.

  Socioeconomic and digital divides in public support for AI-driven welfare decisions: cross-national evidence from 21 countries · 2026 · DOI

• 1. Responsible AI communities should investigate, interrogate, and reflect on cases of AI abandonment occuring prior to deployment to better understand current practices, challenges, and gaps in AI system development. By documenting and analyzing cases of AI abandonment at all stages of development, the community can gain a more complete picture of drivers and blockers towards responsible AI development, including those non-ethics-related. 2. Artifacts and tooling should explicitly increase visibility and outline discussions on non-development or abandonment, particularly where organizations identify that they lack appropriate resources, expertise, or capacity to develop their desired system. 3. Responsible AI development toolkits, frameworks, and artifacts can make more explicit the influence of resources, including costs, technical expertise, and development timelines, on the potential success of other development stages. For example, encouraging organizations to explicitly outline the costs of collecting sufficiently large datasets or maintaining deployed systems can use resource constraints as a lever to discourage development of AI systems that also carry ethical risks or would face development lifecycle challenges. 4. Responsible AI tooling should explicitly encourage organizations to reflect on, outline, and update condi- tions required to appropriately re-visit or resurrect development of abandoned AI systems. 6 Limitations & Future Work The taxonomy presented in §3 is meant to illustrate the diverse levers that can facilitate non-development rather than be exhaustive, and additional factors may emerge in future work. While we aimed to review all available details, analysis of abandoned AIAAIC cases may not capture every single factor that ultimately contributed to AI abandonment, as these databases largely concern deployed systems and external reports may lack visibility into organizational dynamics or resource constraints. Relatedly, data on non-development and abandoned projects is hard to collect as these systems receive fewer resources, are often less formally documented, and can be less salient to practitioners. Our real-world data analyses affirmed this, as very few systems in the AIAAIC repository were reported as abandoned and surveyed practitioners were more inclined to report on continued AI developments rather than abandoned ones. Thus, our work reflects a first step towards this gap by engaging and gathering insights from organizations that abandoned AI development across diverse domains and purposes, particularly in cases prior to system deployment. Future work can continue this approach, encouraging researchers, organizations, and advocacy groups to 1) continue collection, analysis, and sharing of abandoned AI cases across all lifecycle stages, 2) conduct further empirical analysis of incentives and drivers preventing abandonment, and 3) examine the broader impacts of AI system removal.

  To Build or Not to Build? Factors that Lead to Non-Development or Abandonment of AI Systems · 2026 · DOI

• 9.1 Multi-Operator Replication The highest priority for future work is replication of the CharlotteOS methodology across a sample of practitioners with diverse domain expertise, cognitive profiles, and professional contexts. Replication would either confirm the generalizability of the governance architecture or identify the operator-specific factors that currently cannot be separated from the methodology. 9.2 Formal Neurodivergence and AI Augmentation Study The neurodivergent-AI affinity hypothesis warrants a dedicated empirical study comparing multi-model AI orchestration performance across neurodivergent and neurotypical practitioners. Such a study would require validated neurodivergence assessment, standardized task sets across multiple concurrent domains, and output quality metrics that can be evaluated independently of operator self-report. 9.3 Doctoral Pathway This paper is intended to support the operator's doctoral pathway targeting Stanford HAI (AI governance architecture, CharlotteOS methodology, multi-model framework) and the University of Oxford or Cambridge (Tudor genealogical research, GFP methodology, mourning jewelry as historical evidence). The convergence of these two doctoral targets — 'the collision point of Stanford AI research and Cambridge historical scholarship' — is itself an expression of the cross- domain synthesis that CharlotteOS makes operationally sustainable. Formal collaboration with Nitin Aggarwal (Stanford HAI) and Stephen Hall (University of Nottingham, Digital Narrative Care) is being pursued as a near-term priority.

  One Operator, Many Minds: CharlotteOS as a Model for Constitutional Multi-Model AI Orchestration in High-Complexity Knowledge Work · 2026 · DOI

• If yes, could you describe briefly the context and what you found helpful or not? (open question/optional) For each of the following platform functions, please indicate your preferred balance between AI support and human oversight. In this respect, the publicly available documentation provides limited evidence.

  AI4Deliberation_D2.1 – Legal and Ethics Aspects · 2026 · DOI

• To address these limits, future research should focus on empirical validation within Moroccan and North African universities. In this respect, future research is recommended to undertake empirical research, including surveys and interviews with university leaders, faculty, and students, to validate and fine-tune the model in practice.

  Governance of inclusive decision-making in AI adoption · 2026 · DOI

• 149 In fact, it is a form of social influence that scholars have repeatedly tried to distinguish from coercion and persuasion, with mixed results. Although empirical strategies have been developed to counter the widespread surveillance and influence of AI-based technologies,166 what is lacking is a proper reflection and understand- ing of how to effectively protect freedom of thought and conscience once cracks appear on the surface of the “black box” of the human mind, challenging the “boundaries” between what takes place in public and what people consider intimate and private.

  Freedom of Thought, Conscience, and AI · 2026 · DOI

• Specifically, at the level of liable subjects, current norms fail to clarify the standards of duty of care for drivers, automobile manufacturers and algorithm designers; algorithmic black boxes and data barriers cause difficulties in evidence production and responsibility shifting, and the qualification of artificial intelligence as a criminal subject remains controversial.

  Research on the Criminal Imputation of Crimes Involving Level 3 Autonomous Vehicles · 2026 · DOI

• The integration of agentic intelligence into productive and social systems requires a clear and articulated ethical and political vision, embodied in fundamental principles, a regulatory framework, responsible practices at individual, organisational and systemic levels. Europe, through the AI Act, seeks to position itself as a leader in trustworthy and responsible AI. This approach is not without criticism: some argue that excessive regulation undermines competitiveness; others argue that it is still insufficient to address systemic risks. The debate continues, reflecting genuine tensions amongst legitimate viewpoints. The European AI Act, which has come into force progressively from 2024, adopts a risk-based approach: – unacceptable practices (such as governmental social scoring or subliminal manipulation) are prohibited; – high-risk systems (such as those used for credit assessment or the management of critical infrastructure) must meet stringent requirements concerning data quality, documentation, transparency, human supervision and robustness; – other systems are not subject to mandatory requirements (except for certain transparency obligations). The regulatory challenge is to balance risk protection with the promotion of innovation in a domain characterised by extremely dynamic technological evolution, in which generative and agentic models have joined traditional machine-learning approaches. Excessively strict regulation could stifle research and development, whilst self-regulation would leave room for abuses and harm. Moreover, the global nature of AI technologies requires international coordination: fragmented standards could create trade barriers and opportunities for regulatory arbitrage. Ensuring alignment, robustness, scalability, and interoperability requires continuous advances in research and development. However, the most profound challenges are moral and ethical. How should institutions and labour markets be redesigned for productive human–AI collaboration? How can economic benefits be distributed fairly, avoiding concentrations of power? How can systems that are natively opaque be developed to respect democratic values of transparency, accountability and fairness? How can human dignity and autonomy be preserved in cognitive ecosystems increasingly mediated by algorithms? In these complex ‘equations’, the focus must remain not only on the human being but also on the planet as a whole, with an eco-centric vision, since humanity is not separate from, but deeply interdependent with the ecosystem. The debate should not be driven by Big Tech alone; instead, it requires interdisciplinary dialogue.

  From automation to agency: The paradigm of agentic AI across technology, society and the ontology of work · 2026 · DOI

• For GCC IS governance leaders, the five structural antecedents constitute a prospective readiness diagnostic applicable before governance incidents occur. For parent organisation boards, the study demonstrates that uniform enterprise governance standards transferred without adaptation for host-country regulatory environments will reliably produce the readiness gaps documented. For regulators, the multi-jurisdictional complexity documented in P5 argues for coordinated guidance between DPDPA and international regulatory counterparts. The SEC's approach to cybersecurity governance disclosure [23] provides a useful model for regulators seeking to mandate board-level AI governance engagement. enforcement, CERT-In [19], 17 International Journal of Applied Information Systems (IJAIS) – ISSN : 2249-0868 Foundation of Computer Science FCS, Delaware, USA Volume 13– No. 3, June 2026 – www.ijais.org include Three principal limitations bound this study. First, as a multi- site field study, analytical generalisation is to theoretical propositions rather than statistical populations [38]; the five propositions require quantitative testing across larger GCC samples. Second, the cross-sectional design captures configurations at a single point in time. Third, India's DPDPA 2023 enforcement guidance continues to evolve. Future research directions longitudinal study of GCC governance configurations, quantitative survey research testing P1-P5, comparative study across GCC geographies, and dedicated empirical study of agentic AI governance design requirements. 6. CONCLUSION AI governance configurations in GCCs and the relationships between governance levels across different organisational and institutional contexts have received scant attention in academic research. The present study addresses this gap through an in- depth field study of 28 interviews across five GCC organisations, applying the constrained-efficiency framework [1], NIST AI RMF 1.0 [7], and Gioia et al. [6] qualitative methodology.

  Racing Ahead, Governing Behind: An Institutional Analysis of AI Governance Readiness in Global Capability Centres · 2026 · DOI

• The reviewed literature suggests that metaverse-based AI may contribute to educational experiences; however, this interpretation is limited to the scope of the analyzed secondary sources.

  A Study on Artificial Intelligence and the Metaverse · 2026 · DOI

• such as 2.3 Data Collection and Analysis 2.3.1 Data Sources The data sources used in the research are given below:  Academic Articles: Articles related to artificial intelligence and the metaverse scanned in directories such as SCI, SCI-E, and SCOPUS.  Industry Reports: Market analysis reports published by technology companies and working groups.  White Papers: Technical documentation on AR and VR technologies, blockchain infrastructure, and AI algorithms. 2.3.2 Analysis Methods The data were analyzed using thematic analysis and case study methods. The steps followed in this process are given below:  Coding: Data from the literature has been coded to create themes on the intersections between the metaverse and AI.  Thematic Grouping: The coded data were grouped under the main themes, and a structural analysis model was developed.  Evaluation of the Results: The themes obtained were analyzed in depth to make inferences related to the aims of the study. Table 2.

  A Study on Artificial Intelligence and the Metaverse · 2026 · DOI

• While commitment theory has long emphasized the positive influence of individual com- mitment on decision-making [110], research specifically examining CTC remains scarce.

  From IT artifact to AI artifact: an empirical study on the impact of IT identity on user adaptation to AI-driven technological change · 2026 · DOI

• Future studies should focus on international mechanisms for coordinating the regulation of cross-border harms caused by AI chatbots to children. Additionally, future studies are required to reconcile the need for strict safety regulations (Algorithmic Paternalism) and the potential harm of these regulations (the Lobotomy of AI Chatbots and the Denial of Minors Digital Autonomy).

  Duty of Care for Child Protection in AI Chatbots: A Comparative Analysis of EU, US, and Chinese Regulatory Frameworks · 2026 · DOI

• As ML teams mature and roles such as MLOps specialists, AI architects or Responsible 1 3AI and Ethics (2026) 6:339 AI leads become more clearly delineated, future work may investigate how the responsibility for implementing specific Concrete Guidelines or Best Practices is distributed across these roles. While the present study focused on the feasibility of integrating CGs and BPs into real-world ML development workflows, future research may investigate whether their adoption leads to measurable improvements in model-level ethical performance.

  Bridging principles and practice: ethical machine learning in production for developers · 2026 · DOI

• However, both China and the EU currently face a common challenge: while ethical principles are abundant, the institutional density required to translate them into specific procedural obligations, remedial mechanisms, and procurement constraints remains insufficient.

  A Study on the Risks and Ethical Regulation of Artificial Intelligence in Public Decision-Making · 2026 · DOI

• Make override easy (not a multi-step approval process that discourages it) Log every override decision with timestamp and reviewer notes Requirement 3: Transparency and Explainability You must be able to explain to candidates, recruiters, and regulators exactly why a hiring decision was made.⁷ What this means: "The AI scored you 58" is not an explanation "The AI scored you 58 because you mentioned Python 3 times in your resume while the benchmark candidate mentioned it 7 times, and you had a 2-year gap in employment while strong candidates had no gaps" is closer to an explanation Regulation specifically requires "meaningful information about the logic" behind decisions.⁸ This isn't marketing language. It's regulatory language. Compliance action: Implement explainability logging that captures the specific evidence for each decision Make explainability reports accessible to candidates (many EU jurisdictions require notification) Document the reasoning methodology—how does your system weight factors? Requirement 4: Bias Testing and Disparate Impact Analysis You must test whether your hiring AI has a disparate impact on protected groups. What this means: Organizations must conduct statistical analysis showing whether their AI system rejects candidates from protected groups (race, gender, age, disability, etc.) at different rates than other groups. Under US Equal Employment Opportunity rules, a difference of more than ~5% in adverse impact is considered evidence of discrimination.⁹ The EU standard is similarly rigorous. Compliance action: Run disparate impact analysis on your hiring AI outcomes across demographic groups Document the results (even if they show no disparate impact) If disparate impact is found, either correct the system or document why you're accepting the impact as justified Conduct this analysis annually and document findings Requirement 5: Record-Keeping and Audit Trails Every hiring decision must be logged in a way that allows complete reconstruction of the decision-making process. What this means: Who made the decision? When was it made? What system version was used? (If you update your AI model, each candidate's decision is traceable to the specific model version that scored them) What data was considered? Who reviewed it? What was the outcome? (Hired, rejected, advanced) What actually happened? (If hired: 30-day performance, 90-day performance, 1- year retention) This is audit-grade logging. It's not optional.

  The CHRO's 18-Month Compliance Window: How to Be Audit-Ready by December 2, 2027 · 2026 · DOI

• Yet, existing empirical approaches often treat norms as targets for alignment or replication, implicitly assuming equivalence between human subjects and AI agents and leaving collective normative dynamics insufficiently examined.

  Normative Common Ground Replication (NormCoRe): Replication-by-Translation for Studying Norms in Multi-Agent AI · 2026 · DOI

• to optimize rollout phases. AI insights are fed into our PM dashboards, but senior leadership validates the final calls. AI assists with prioritization, not replacement. 2. Data silos and organizational inertia. We broke down silos by establishing cross- functional data teams and ran change management programs to support cultural adaptation 3. AI optimizes our network traffic, enhances customer service with chatbots, and automates billing processes. It’s also helping us transition to 5G by predicting capacity needs. Interviewee 7: Senior Project Consultant, Global Engineering Firm: 1. We’ve adopted AI for resource scheduling, and site monitoring using drones, construction modeling. AI enhances accuracy and minimizes rework in design- build projects.

  Artificial Intelligence–Driven Strategies for Enhancing Digital Transformation and Project Risk Management · 2026 · DOI

• This study offers valuable perceptions into AIbased digital transformation and project risk management; various the breadth and depth of the research. One main limitation is the small sample size; only 3 interviews have been conducted due to time restraints and limited access to experts with significant experience. A more diverse participant pool, including individuals from various areas like forced https://doi.org/10.54489/ijbas.v6i1.599 Published by: GAFTIM, https://gaftim.com F. Al Jaberi et al. International Journal of Business Analytics and Security (IJBAS) 6(1) -2026- 88 logistics, finance, and healthcare, might have enhanced the findings and included wider insights. and results Figure 7: Limitations of this Research (Source: Self-made) Another limitation is the lack of longitudinal data. This study takes experiences at a specific point in time, which might not fully represent current practices of AI long-term implementation. Understanding organizations over time might provide a more complete knowledge of AI’s sustained effect on project risk management. Similarly, for resource constraints, the study does not involve workshop or group discussions, which may have permitted a better evaluation of collaborative stakeholder dynamics. Additionally, more widespread case study comparisons between UAE-based and global organizations might have included vital crosscultural dimensions. This research does not focus on the lack of quantitative insights, which offers more in-depth and statistical data regarding the topic, increasing generalizability and offering measurable results. Enhancing access to AI tool performance metrics and proprietary project data might enhance the depth and reliability of the evaluation. viewpoints and • Recommendations Based on the literature review, international best practices, evolving trends, and interview findings, the following targeted suggestions are offered for implementing AI into project risk management and digital transformation frameworks.

  Artificial Intelligence–Driven Strategies for Enhancing Digital Transformation and Project Risk Management · 2026 · DOI

• Third, the analysis focused on enterprise-level responses and did not account for consumer-side factors or informal technological diffusion. First, the data used were limited to a single Eurostat survey, which may not fully capture the dynamic and sector-specific nature of AI adoption, especially in emerging industries.

  Statistics on the use of AI technologies in the member states of the EU · 2026 · DOI

• Future research should focus on the ambivalence revealed in this article. A possible explanation for such a result is that older generations, regardless of digital literacy, may have a higher level of institutional trust in science as such, which acts as compensation for the perceived risks of the tool itself, what should be examined in greater detail in future studies.

  Predictors of Affirmative Attitudes toward the Use of Artificial Intelligence in Science within a Mertonian Ethos Framework · 2026 · DOI

• This study is limited to a selected set of international AI ethics reports published between 2019 and 2023 by major intergovernmental and regional organizations. National-level policies, sector-specific regulations, and corporate AI ethics frameworks fall outside the scope of the analysis. For instance, national frameworks such as Singapore's Model AI Governance Framework, China's New Generation AI Development Plan, or Brazil's AI Strategy reflect region-specific cultural values, economic priorities, and regulatory traditions that may differ substantially from the international frameworks analyzed here. Similarly, sector-specific guidelines developed by professional associations (e.g., medical AI ethics by healthcare regulatory bodies) or corporate AI principles (e.g., Microsoft's Responsible AI Standards, Google's AI Principles) embody organizational and industry- specific interpretations of ethical commitments that are not captured in this study. As a result, certain contextual variations in how AI governance is interpreted and implemented at national, sectoral, or organizational levels may not be fully captured. Moreover, these documents precede several significant regulatory developments, most notably the adoption of the European Union Artificial Intelligence Act in 2024, which marks a transition from predominantly voluntary ethical guidance toward legally binding regulatory obligations. Nevertheless, the analytical contribution of the present study does not lie in assessing the effectiveness of these newer regulatory instruments, but in systematically examining how foundational ethical principles were initially articulated and embedded within distinct institutional and governance logics.

  ARTIFICIAL INTELLIGENCE AND ETHICS: A GLOBAL PERSPECTIVE · 2026 · DOI

• This study integrates structural modeling and qualitative inquiry to advance under- standing of how AI-assisted tool use is associated with students’ regulation, innova- tion, and ethical judgment in design-related higher education. Nevertheless, several limitations should be noted. First, the sample was limited to students enrolled in design-related programs, which restricts the generalizability of the findings. Although the model offers useful insights into how AI engagement is associated with self-regulated learning, inno- vative behavior, critical thinking, and creative integrity, these relationships should not be assumed to operate identically across all disciplines. Future studies should examine the proposed framework in other academic domains and cultural contexts to determine whether similar behavioral mechanisms emerge under different learning conditions. Second, the study relied on cross-sectional SEM data and retrospective interviews, which limited its ability to capture real-time regulatory, emotional, and behavioral processes. Students’ interactions with AI are likely to evolve across tasks, feedback cycles, and stages of learning. Future research could employ longitudinal designs, learning analytics, process-tracing methods, or digital trace data to examine how 1 3Reframing human–AI collaboration in higher education: behavioral… self-regulation and innovative behavior develop dynamically in AI-mediated learn- ing environments. Third, the model focused primarily on self-regulated learning and innovative behavior as mediating mechanisms. Although these constructs were theoretically grounded and empirically supported, other variables may also shape the relation- ship between AI use and higher-order outcomes. Future studies could incorporate constructs such as AI literacy, digital trust, epistemic vigilance, design thinking, or ethical sensitivity to provide a more comprehensive account of the human–AI–ethics ecosystem. Fourth, the contextual scope of this study was limited to higher education settings. The ethical and practical implications of AI-assisted creativity may differ substan- tially in professional and workplace environments, where creative outputs are subject to client expectations, commercial value, legal accountability, intellectual property regulations, and reputational consequences. In such high-stakes contexts, creative integrity may no longer function only as an internal reflective capacity but also as an externally evaluated professional responsibility. Similarly, self-regulation and innovative behavior may shift under workplace pressures, where efficiency, original- ity, authorship, and ethical compliance must be negotiated simultaneously. Future research should therefore test the model in commercial design studios, professional creative industries, and other workplace settings to examine whether the pathways identified in this study remain stable or are reshaped by institutional, legal, and mar- ket constraints. By acknowledging these limitations, future research can refine the theoretical pre- cision, empirical scope, and practical relevance of the proposed framework. Such efforts may help clarify the conditions under which AI integration supports, rather than undermines, critical, innovative, and ethically grounded forms of creativity.

  Reframing human–AI collaboration in higher education: behavioral pathways from generative AI to critical thinking and ethical creativity · 2026 · DOI

Most-cited papers in Ethics and Social Impacts of AI

• Artificial Intelligence and Management: The Automation–Augmentation Paradox · Academy of Management Review · 2021 · 1,776 citations
• Generative AI and the future of education: Ragnarök or reformation? A paradoxical perspective from management educators · The International Journal of Management Education · 2023 · 860 citations
• The ethics of AI in health care: A mapping review · Social Science & Medicine · 2020 · 844 citations
• Transparency and the Black Box Problem: Why We Do Not Trust AI · Philosophy & Technology · 2021 · 643 citations
• Rising with the machines: A sociotechnical framework for bringing artificial intelligence into the organization · Journal of Business Research · 2020 · 598 citations
• What influences algorithmic decision-making? A systematic literature review on algorithm aversion · Technological Forecasting and Social Change · 2021 · 508 citations
• Paradoxes of artificial intelligence in consumer markets: Ethical challenges and opportunities · Journal of Business Research · 2020 · 441 citations
• Accountable Artificial Intelligence: Holding Algorithms to Account · Public Administration Review · 2020 · 430 citations
• Artificial intelligence, systemic risks, and sustainability · Technology in Society · 2021 · 400 citations
• When combinations of humans and AI are useful: A systematic review and meta-analysis · Nature Human Behaviour · 2024 · 387 citations

Most recent work

• Paper 4 — Role Society: Human Discretion, Accountability, and Non-Identifiable Role Coordination in the AGI Era · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Paper 7 — AGI Output Governance: Candidate Outputs, Human Discretion, and Evidentiary Boundaries in the AGI Era · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Paper 8 — HTS-Based Evidence Sealing in the AGI Era: Preserving Output History, Human Discretion, and Evidentiary References · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Paper 11 - Structural Operating Framework for the AI+AGI Era · Zenodo (CERN European Organization for Nuclear Research) · 2026
• SΔϕ-57 — Lent Thought: Thought Ownership, World-Binding, and Cost Attribution in Human-AI Reasoning · Zenodo (CERN European Organization for Nuclear Research) · 2026
• SΔϕ-55 — Transition Governance Alignment Index: Minimal Quantification of Alignment after SΔϕ-42 · Zenodo (CERN European Organization for Nuclear Research) · 2026
• Paper 23 — Human-Centered Structured Verification: From Claim Decomposition to Research, Planning, and Social Consensus in the AI+AGI Era · Zenodo (CERN European Organization for Nuclear Research) · 2026
• SΔϕ-48 — AI Drift as Transition Governance Instability: Authority, Refusal, World-Binding, Editability, and Re-entry (v1.1, AI-Readable Package) · Zenodo (CERN European Organization for Nuclear Research) · 2026
• SΔϕ-49 — AI Poverty as Failure Non-Reentry Resource Deficit: Context, Tool, Authority, Verification, Feedback, Editability, and Re-entry Path Deficits (v1.1, AI-Readable Package) · Zenodo (CERN European Organization for Nuclear Research) · 2026
• The Representation Economy: Canonical Framework v1.0 · Zenodo (CERN European Organization for Nuclear Research) · 2026

Related topics in Social Sciences