[IGF-DAIG] DRAFT DAIG CfP AI Safety & Security

Tue Oct 14 02:39:14 CEST 2025

Dear all,
I hope this message finds you well!
Together with some DAIG friends, we discussed the possibility to dedicate the next DAIG Call for Papers to the issue of AI Safety and Security. These are two rather pressing issues on which many of us are developing research. Of course, this is only a suggestion, and other alternative options are more than welcome.
Please find here  https://pad.codigosur.org/DRAFT_DAIG_CfP_AI_Safety_%26_Security and below the draft Call for Papers for the 2026 DAIG Annual Report. The goal of this edition would be to explore both the safety dimension, i.e. understanding and preventing unintended harm from non-malicious failures, and the (cyber)security dimension, i.e. protecting AI systems from adversarial and malicious threats.
The document builds on our previous reports and proposes a selection of potential AI safety and security policy issues, reflecting the growing importance of these topics in global AI governance discussions.
We kindly invite all DAIG members to review the draft and share feedback, comments, or proposed revisions (or alternative suggestions) by 30 October 2025. Your input will be instrumental in ensuring the call accurately reflects the Coalition’s priorities and supports inclusive, high-quality contributions.
Please send your comments either by replying to this email or adding them directly on the pad https://pad.codigosur.org/DRAFT_DAIG_CfP_AI_Safety_%26_Security.
Thank you very much and kind regards
Luca

DRAFT CfP:
IGF Coalition on Data and Artificial Intelligence Governance

Call for Papers onAI Safety and Security

Background

The IGF Coalition on Data and Artificial Intelligence Governance (DAIG) is a multistakeholder group established under the auspices of the United Nations Internet Governance Forum (IGF). The Coalition fosters dialogue on data and AI governance, encourages analysis of successful and unsuccessful practices, and promotes collaboration to identify solutions that advance sustainable and effective AI governance.

To this end, the DAIG promotes collective studies and multistakeholder interactions to gather evidence, critically analyse regulatory and institutional arrangements, and propose policy updates in AI governance. The Coalition acts as a hub connecting global UN IGF discussions with regional and local initiatives, with a particular focus on debates and experiences from the Global South.

The 2025 Annual Report of the Coalition will focus on AI Safety and Security, highlighting both technical and policy challenges and showcasing experiences from diverse contexts to inform the IGF 2026 discussions.

Call
The IGF Coalition on Data and Artificial Intelligence Governance (DAIG), under the auspices of the United Nations Internet Governance Forum, invites researchers, policymakers, and practitioners to submit short papers analysing issues related to AI Safety and Security.

Submissions may explore theoretical, empirical, technical, or policy-oriented perspectives.

Suggested Topics
The topics below are indicative and not exhaustive. Papers may address other relevant dimensions of AI safety and security.
AI Safety
·       System robustness and reliability: Approaches to ensure consistent AI performance across environments, addressing vulnerabilities due to software bugs, poor input validation, or unstable model behavior.
·       Specification and assurance: Methods to align AI objectives with human intent, including formal verification, interpretability, and human-in-the-loop mechanisms.
·       Bias and fairness: Detecting and mitigating bias that can lead to harmful or discriminatory outcomes, such as misclassification in medical or hiring applications.
·       Safety in autonomous systems: Ensuring reliable perception, decision-making, and control in safety-critical applications (e.g., autonomous vehicles, healthcare robotics).
·       Safe deployment of large models: Techniques to ensure large language models (LLMs) and foundation models produce accurate, harmless, and unbiased outputs.
·       Testing, evaluation, and certification: Frameworks for AI safety auditing, benchmarking, and lifecycle assurance.
·       AI safety governance and policy: Legal, ethical, and institutional measures for ensuring safe AI deployment, including accountability and oversight mechanisms.
AI Security
·       Adversarial attacks and defenses: Techniques to detect and mitigate attacks that manipulate model inputs or outputs.
·       Data poisoning and model corruption: Safeguarding training data against malicious modification that induces harmful or unpredictable behavior.
·       Model extraction and inversion attacks: Preventing unauthorized access to proprietary model parameters or sensitive training data through inference or querying.
·       Prompt injection and prompt-based exploits: Defending LLMs and chat-based AI from manipulation to generate harmful, misleading, or confidential outputs.
·       Supply chain security: Securing datasets, pre-trained models, and software dependencies used in AI system development.
·       AI-enabled cyberattacks: Understanding and mitigating how AI can be weaponized in phishing, misinformation, and automated hacking.
·       Security auditing, monitoring, and incident response: Frameworks and best practices for ensuring resilience of AI systems to attacks.
·       AI security policy and international cooperation: Governance models, risk management frameworks, and global coordination for secure AI deployment.
Cross-cutting Topics
·       Interplay between AI safety, AI security, and AI ethics.
·       Policy design for global AI safety and security governance.
·       Standards and certification for trustworthy AI.
·       Capacity building and technology transfer for AI safety and security in the Global South.

Submission Guidelines
Research papers — including analytical, theoretical, position papers, or case studies — will be considered for inclusion in the 2025 DAIG Report, even if previously published.

• Length: 2,500–5,000 words
• Authorship: Up to 3 co-authors
• Language: English or Spanish
• Style: APA 6, using footnotes (not endnotes)
• Format: Microsoft Word or OpenDocument Text

Submissions must include:
1. Title
2. Short abstract (200 words)
3. Keywords
4. Draft paper
5. In the body of the email: author’s name, affiliation, and short biographical note

Deadline: 28 February 2026

Submissions should be sent to luca.belli [at] fgv.br and walter.gaspar [at] fgv.br, with “Paper Submission DAIG 2025” as the subject line.

Authors will be notified within approximately seven days from the deadline regarding the status of their contributions.

All submissions will be subject to peer review. Each author submitting a paper will be asked to review another submission within 10 days, assessing (1) novelty, (2) theoretical soundness, and (3) quality of presentation. Authors will have the opportunity to revise their contributions based on peer comments.

Selected papers will be published in the 2025 Report of the IGF Coalition on Data and Artificial Intelligence Governance, under open-access Creative Commons licenses.

Authors will also be invited (at their own expense) to present their work at the annual DAIG meeting at the United Nations Internet Governance Forum (IGF), to be held in 2026 (venue and dates to be announced).

[http://www.fgv.br/mailing/2018/direito_rio/assinatura/fgv_direito_rio.jpg]

Luca Belli, PhD

Professor | FGV Law Schoool

Coordinator | Center for Technology & Society

[http://www.fgv.br/mailing/2018/direito_rio/assinatura/fgv_tel.jpg]+55 21 3799 5763 t at 1lucabelli<https://twitter.com/1lucabelli>

[http://www.fgv.br/mailing/2018/direito_rio/assinatura/fgv_direito_rio_map.jpg]Praia de Botafogo, 190 13º andar

Botafogo - Rio de Janeiro, RJ - CEP: 22250-900

[http://www.fgv.br/mailing/2018/direito_rio/assinatura/fgv_ass_url.jpg]personal www.lucabelli.net<https://lucabelli.net/>

[http://www.fgv.br/mailing/2018/direito_rio/assinatura/fgv_ass_url.jpg]CTS-FGV direitorio.fgv.br/cts<https://direitorio.fgv.br/cts>

[http://www.fgv.br/mailing/2018/direito_rio/assinatura/fgv_ass_url.jpg]CPDP LatAm www.cpdp.lat<http://www.CPDP.lat>

[http://www.fgv.br/mailing/2018/direito_rio/assinatura/fgv_ass_url.jpg]cyberBRICS www.cyberbrics.info<http://www.cyberbrics.info/>

CONFIDENTIALITY NOTICE
This message, as well as any attached document, may contain personal data and information that is confidential and privileged and is intended only for the use of the addressee named above. If you are not the intended recipient, you are hereby notified that any disclosure, copying or distribution of this email or attached documents, or taking any action in reliance on the contents of this message or its attachments is strictly prohibited and may be unlawful. Please contact the sender if you believe you have received this email by mistake.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.ghserv.net/pipermail/igf-daig/attachments/20251014/5f91cfb0/attachment-0001.htm>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Outlook-http___www.png
Type: image/png
Size: 14321 bytes
Desc: Outlook-http___www.png
URL: <http://lists.ghserv.net/pipermail/igf-daig/attachments/20251014/5f91cfb0/attachment-0007.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Outlook-http___www.png
Type: image/png
Size: 365 bytes
Desc: Outlook-http___www.png
URL: <http://lists.ghserv.net/pipermail/igf-daig/attachments/20251014/5f91cfb0/attachment-0008.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Outlook-http___www.png
Type: image/png
Size: 566 bytes
Desc: Outlook-http___www.png
URL: <http://lists.ghserv.net/pipermail/igf-daig/attachments/20251014/5f91cfb0/attachment-0009.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Outlook-http___www.png
Type: image/png
Size: 668 bytes
Desc: Outlook-http___www.png
URL: <http://lists.ghserv.net/pipermail/igf-daig/attachments/20251014/5f91cfb0/attachment-0010.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Outlook-http___www.png
Type: image/png
Size: 668 bytes
Desc: Outlook-http___www.png
URL: <http://lists.ghserv.net/pipermail/igf-daig/attachments/20251014/5f91cfb0/attachment-0011.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Outlook-http___www.png
Type: image/png
Size: 668 bytes
Desc: Outlook-http___www.png
URL: <http://lists.ghserv.net/pipermail/igf-daig/attachments/20251014/5f91cfb0/attachment-0012.png>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: Outlook-http___www.png
Type: image/png
Size: 668 bytes
Desc: Outlook-http___www.png
URL: <http://lists.ghserv.net/pipermail/igf-daig/attachments/20251014/5f91cfb0/attachment-0013.png>