ArabicNLP 2026 — Reviewer Ethics Guide
This guide provides practical instructions for reviewers to identify and report ethical issues in submissions.
Source: https://aclrollingreview.org/ethicsreviewertutorial
1. Scope of Ethics Review
Ethics review is not a technical review.
Focus only on:
Whether the submission raises substantial ethical concerns
Whether risks are acknowledged and mitigated
Do NOT evaluate:
model performance
experimental design
writing quality
novelty (unless tied to ethical risk)
A substantial ethical issue is one that creates a meaningful risk of harm beyond typical NLP research norms.
2. What Reviewers Should Check
2.1 Potential Harm and Misuse
☐ Does the work enable harm (e.g., surveillance, profiling, misinformation)?
☐ Are risks to individuals, groups, or society identified?
☐ Are mitigation strategies discussed?
2.2 Bias and Fairness
☐ Does the work introduce or amplify bias?
☐ Are impacts on specific groups (e.g., dialects, communities) considered?
☐ Are fairness limitations acknowledged?
2.3 Data Ethics and Privacy
☐ Is the data source clearly described?
☐ Is there evidence of consent, licensing, or authorization?
☐ Are privacy risks (e.g., re-identification) addressed?
2.4 Use of Data, Code, and Resources
☐ Are datasets and tools properly cited?
☐ Is there any indication of unauthorized use?
☐ Are licenses respected?
2.5 Human Subjects and Labor
☐ Are annotators or participants treated ethically?
☐ Are risks or burdens on participants acknowledged?
2.6 Research Integrity
☐ Any signs of plagiarism or misrepresentation?
☐ Any misleading claims or omitted results?
☐ Any attempt to compromise anonymity?
2.7 Harmful or Sensitive Content
☐ Is harmful content (e.g., hate speech) necessary for the research?
☐ Is it properly contextualized?
☐ Are warnings or safeguards included where appropriate?
2.8 ArabicNLP-Specific Considerations
☐ Are dialects and linguistic varieties treated fairly?
☐ Are communities represented accurately?
☐ Are sociopolitical sensitivities handled responsibly?
3. How to Write an Ethics Review
When raising an issue:
Clearly describe the concern
Link it to a recognized ethical principle (e.g., harm, fairness, privacy)
Suggest concrete ways to address it
Good review structure:
Identify the issue
Explain why it is ethically problematic
Suggest mitigation or improvement
Avoid:
vague statements without justification
relying on country-specific regulations
purely technical criticism
4. Reviewer Conduct
Reviewers must:
maintain confidentiality of submissions
not use content for personal advantage
not upload papers to external AI tools
avoid harassment, bias, or unprofessional language
Reviewers must NOT:
attempt to identify authors
collude or manipulate the review process
request unnecessary citations to their own work
5. Use of Generative AI in Reviewing
☐ Do not use AI tools to generate reviews
☐ Do not upload submissions to non-private tools
☐ Limited use for language polishing is acceptable
6. When to Flag a Paper
Flag for ethics review if:
risks are significant and unaddressed
data use appears unethical or unauthorized
harm potential is high or unclear
there are signs of misconduct
7. Key Principle
Ethics review is about risk awareness and responsibility, not punishment.
The goal is to ensure that research is:
responsible
transparent
aligned with community standards
For questions or uncertain cases, contact the Ethics Chairs : arabicnlp-ethics-chairs@sigarab.org