Call for Papers:

Topical Collection on AI Agents: Ethics, Safety, and Governance
Springer Journal AI and Ethics
https://link.springer.com/journal/43681/updates/27832482

Guest Editors:
Rebecca L. Johnson (The University of Sydney)
Aida Mostafazadeh Davani (Google Research)
Leslye Denisse Dias Duran (Institute for Ethics in AI, University of Oxford)
Tricia A. Griffin (Maastricht University)
Lorenn P. Ruster (The Australian National University)
Jennifer Munt (The Australian National University)

Deadline for submissions: May 31, 2026

Artificial intelligence is undergoing a structural shift. Systems once framed as tools that answer are becoming systems that act—planning, reasoning, executing multi-step procedures, coordinating with other agents, and adapting with limited oversight. These “AI agents” can use tools, write and run code, and pursue open-ended goals across digital and physical environments. Their rapid integration into homes, workplaces, public services, and critical infrastructure expands both capability and risk, raising ethical, safety, and governance questions that exceed the frameworks developed for static or conversational models. Addressing this transition requires interdisciplinary work that combines technical insight with philosophical and societal analysis.

For this collection, an AI agent is an AI system that can select, sequence, and execute actions affecting digital or physical environments, with varying levels of autonomy. This includes assistants, tool-using LLM pipelines, and both embodied and software agents capable of modifying their own plans or engaging with other agents.

Existing ethical and technical frameworks, developed for static or generative models, are inadequate for systems that act over time, learn from feedback, and participate in sociotechnical environments. Agents are rapidly entering homes, workplaces, public administration, finance, healthcare, and defence. Misaligned objectives or inappropriate deployment may produce wide-ranging harms.

Scholarship has examined agentic capability, responsibility, autonomy, duty of care, and controllability [1–14]. These studies highlight the need for ethical and governance frameworks that are empirically grounded, philosophically coherent, and suited to systems that act, adapt, and learn within institutions.

This topical collection invites conceptual, empirical, and applied research on the ethical, safety, and governance questions raised by AI agents. Submissions from philosophy, AI safety, engineering, law, HCI, policy, and the social sciences are welcome. Core terms—agency, autonomy, trustworthiness, responsibility—remain contested. The aim is to clarify these concepts and build methodological foundations for understanding and governing AI agents as components of human systems.

Static benchmarks are insufficient for evaluating dynamic, adaptive systems. Safety and risk analysis must address contexts in which agents co-construct norms with humans and institutions, and where plural values and emergent behaviours challenge standard metrics.

The ten challenge areas below illustrate current concerns and invite further exploration.

Ten Ethical Challenge Areas for Agentic AI

1. Human–Agent Relationships, Human-Agent Collaboration, and Anthropomorphism: As AI agents become increasingly fluent, people may overattribute understanding, intention, or care. How should we evaluate the moral significance of empathy or attachment toward entities that simulate but do not possess understanding? What forms of collaboration emerge when agents exhibit fluent behaviour without genuine comprehension? How can design support effective collaboration while mitigating anthropomorphic misreadings?
2. Cross-Cultural and Plural Values: Global deployment amplifies some moral frameworks while marginalising others. To what extent can moral pluralism be computationally represented without collapsing into relativism or moral imperialism? What methods can calibrate agents’ moral profiles against empirical human data across societies and worldviews?
3. Multi-Agent Ecosystems and Emergent Behaviour: AI agents increasingly coordinate, compete, or create new agents. Can moral or political philosophy offer models for governance among non-human actors? What oversight mechanisms can prevent collusion or uncontrolled emergent dynamics in multi-agent environments?
4. Evaluation and Assurance in the Wild: Static benchmarks fail to capture dynamic, real-world behaviour. What would it mean for a system to be ethically reliable rather than merely technically robust? How can longitudinal and contextual evaluation track decision-making “in the wild” and assess ethical performance across environments?
5. Value Alignment under Autonomy: AI agents plan and act independently, decomposing goals in ways that may diverge from human intentions. What conception of moral agency should underpin attempts to align systems capable of self-directed planning? How can alignment be maintained when agents self-plan or self-modify without continuous oversight, within plural and sometimes conflicting human values?
6. Responsibility, Accountability and Liability: Systems incorporating AI agents distribute responsibility across designers, deployers, and users. Can accountability meaningfully apply to artefacts lacking intention, or should it be reconceived as a distributed ethical relation? When an autonomous agent causes harm, how should we distinguish between responsibilities assigned in advance and accountability or liability assessed after the fact?
7. Transparency and Interpretability in Action: AI agents execute multi-step plans, call APIs, and interact with external systems. What counts as a reason or justification when behaviour emerges from procedural learning rather than deliberation? How can decision pathways and action traces be made intelligible to humans without undermining agent performance?
8. Adaptive Autonomy and Goal Drift: AI agents dynamically decompose and redefine goals, producing subtle misalignments. Does moral responsibility require stability of intention? What safeguards can detect and correct goal or value drift while balancing initiative with containment?
9. Institutional and Societal Integration: AI agents are entering workplaces, governance systems, and public services. What does it mean for a sociotechnical system to exhibit institutional “agency,” and how should that shape responsibility? How can organisations ethically delegate decision-making to agents while preserving oversight and public accountability?
10. Security, Misuse, and Malicious Agents: AI agents can be repurposed or exploited, particularly when able to write or execute code. Do self-replicating or malicious agents challenge conventional boundaries between artefacts and actors? What preventive or containment measures can mitigate misuse, replication, or adversarial adaptation?

By examining how agency, autonomy, and responsibility unfold in practice, this topical collection seeks to consolidate and advance the emerging field of Agentic AI Ethics. We invite contributions that clarify core concepts, analyse real-world deployments, propose evaluative and governance frameworks, and explore ethical questions that arise when AI systems act within human institutions. Our aim is to build a shared foundation for understanding and governing AI agents—one that is rigorous, interdisciplinary, and responsive to the complex environments in which these systems will increasingly operate.

Guest Editors

Rebecca L. Johnson (Lead Guest Editor), PhD in AI Ethics, The University of Sydney, Australia, rebecca.johnson@sydney.edu.au
Aida Mostafazadeh Davani, PhD, Senior Research Scientist at Google Research, Portland, USA, aidamd@google.com
Leslye Denisse Dias Duran, PhD, Institute for Ethics in AI, Faculty of Philosophy, University of Oxford, lessdias@gmail.com
Tricia A. Griffin, PhD, Maastricht University, The Netherlands, griffin.triciaann@gmail.com
Lorenn P. Ruster, PhD, School of Cybernetics, The Australian National University, Australia, lorenn.ruster@anu.edu.au
Jennifer Munt, PhD, The Australian National University, Australia, jennifer.munt@anu.edu.au

Manuscript Submission Deadline

31st May 2026

Submission

Submissions should be original papers and should not be under consideration for publication elsewhere. Extended versions of high quality conference papers that are already published at relevant venues may also be considered as long as the additional contribution is substantial (at least 30% of new content).

Authors must follow the formatting and submission instructions of the AI and Ethics journal at https://www.springer.com/journal/43681.

During the first step in the submission system Snapp, please select the appropriate article type. In further steps, please confirm that your submission belongs to a collection and choose from the drop-down menu the appropriate collection title.

Review Process:

Guest Editors (GEs) will follow the review guidelines, policy and procedure of AI and Ethics (AIET). Guest editors are responsible for tracking and managing the review progress, will pre-assess the quality of submitted journal papers. If the submitted papers are not up to the standard, then the submitted papers will be desk rejected by the GEs. The submitted papers with good quality will move to the next stage in which GEs will source at least two independent expert reviewers from their own networks and/or AIET’s reviewers database to robustly review the papers. After receiving the review reports, the GEs will assess them before making recommendations to the Editor in Chief.

References

[1]  Iason Gabriel, Geoff Keeling, Arianna Manzini, and James Evans. 2025. We need a new ethics for a world of AI agents. arXiv preprint arXiv:2509.10289 (2025).

[2]  Iason Gabriel, Arianna Manzini, Geoff Keeling, Lisa Anne Hendricks, Verena Rieser, Hasan Iqbal, Nenad Tomašev, Ira Ktena, Zachary Kenton, and Mikel Rodriguez. 2024. The ethics of advanced ai assistants. arXiv preprint arXiv:2404.16244 (2024).

[3]  Luke R. Hansen. 2023. On the existence of robot zombies and our ethical obligations to AI systems. Journal of Social Computing 4, 4 (2023), 270–274.

[4]  Sayash Kapoor, Noam Kolt, and Seth Lazar. 2025. Build Agent Advocates, Not Platform Agents. arXiv preprint arXiv:2505.04345 (2025).

[5]  Atoosa Kasirzadeh and Iason Gabriel. 2025. Characterizing ai agents for alignment and governance. arXiv preprint arXiv:2504.21848 (2025).

[6]  Sebastian Krügel, Andreas Ostermaier, and Matthias Uhl. 2022. Zombies in the loop? Humans trust untrustworthy AI-advisors for ethical decisions. Philosophy & Technology 35, 1 (2022), 17.

[7]  Margaret Mitchell, Avijit Ghosh, Alexandra Sasha Luccioni, and Giada Pistilli. 2025. Fully autonomous ai agents should not be developed. arXiv preprint arXiv:2502.02649 (2025).

[8]  Joon Sung Park, Joseph C. O’Brien, Carrie J. Cai, Meredith Ringel Morris, Percy Liang, and Michael S. Bernstein. 2023. Generative agents: Interactive simulacra of human behavior. arXiv preprint arXiv:2304.03442 (2023).

[9]  Roberto Redaelli. 2025. Intentionality gap and preter-intentionality in generative artificial intelligence. AI & SOCIETY 40, 4 (2025), 2525–2532.

[10]  Abootaleb Safdari. 2025. An outline of enactive relationalism in the philosophy and ethics of robotics. Philosophy & Technology 38, 2 (2025), 52.

[11]  Adam Safron, Inês Hipólito, and Andy Clark. 2023. Bio AI-from embodied cognition to enactive robotics. Frontiers Media SA.

[12]  Julien da Silva. 2023. Looking the Part: The AI Zombie Problem and the Anti-Turing Test. In Canadian AI, 2023.

[13]  Shannon Vallor. 2024. The AI mirror: How to reclaim our humanity in an age of machine thinking. Oxford University Press.

[14]  Carissa Véliz. 2021. Moral zombies: why algorithms are not moral agents. AI & society 36, 2 (2021), 487–497.