Artificial Intelligence
A Risk-Assessment Framework for Responsible AI Use in Research Publishing
Cureus supports the responsible and transparent use of artificial intelligence where it strengthens research quality, upholds editorial independence, and preserves the integrity of the scientific and scholarly record.
These policies set out a risk-assessment framework governing how AI may be used across the research and publishing lifecycle, applied consistently to authors, peer reviewers, and editors, while recognizing their distinct roles and responsibilities.
AI is treated as a supporting technology. Scholarly judgment, accountability, and responsibility always remain human.
Core Expectations
Across all editorial roles and activities:
- Human accountability is non-transferable
- Accountability for scholarly content, evaluation, and editorial decisions cannot be delegated to AI systems.
- AI may support, but must not replace, scholarly judgment
- AI can assist clarity, efficiency, and exploration, but must not determine conclusions, evaluations, or decisions.
- Transparency creates trust and confidence
- Transparent declaration of AI builds trust and confidence, and removes ambiguity about how AI might have been used.
- Confidentiality and data protection are mandatory
- Manuscripts, peer review reports, and sensitive data must not be shared with unsecured or public AI systems.
Risk-Assessment Framework for AI Use and Declaration
In support of the policy, the following framework provides guidance on how to use AI safely. Use of AI is governed according to risk level, not by tool type. The same framework applies across all roles: authors, reviewers and editors.
| Risk Assessment Level | Green - Permitted: Assistive AI use (low risk) |
|---|---|
| Summary | AI use that supports expression, organisation, or efficiency without influencing scientific, scholarly or evaluative judgment. |
| Expectations |
|
| Examples |
|
| Compliance and Disclosure | Permitted. Disclosure enhances trust and transparency and clearly demonstrates human accountability. |
| Risk Assessment Level | Amber - Exercise caution: Evaluative or interpretive AI use (requires caution and care) |
|---|---|
| Summary | AI use that may influence interpretation, framing, emphasis, or evaluative judgment, but remains under human control. |
| Expectations |
|
| Examples |
|
| Compliance and Disclosure | Permitted with human oversight, verification, and transparency through disclosure. If AI materially influences evaluative judgment, accountability must remain clearly human-led and defensible. |
| Risk Assessment Level | Red - Not permitted: AI replacing scholarly judgment or lacking transparency |
|---|---|
| Summary | AI use that is opaque or replaces accountable human contribution, generates unverifiable outputs, or compromises confidentiality or integrity. |
| Expectations |
|
| Examples |
|
| Compliance and Disclosure | Not permitted |
Role-Specific Responsibilities
The same risk framework applies to all roles, with different responsibilities.
Authors
- Remain fully accountable for originality, accuracy, and integrity
- Disclose AI use for trust and transparency
Peer Reviewers
- Remain responsible for independent, expert evaluation
- May use AI to support peer review but must not upload manuscript content to unsecured or public AI tools
- Disclose AI use for trust and transparency
- Must not delegate critique or judgment to AI systems
Editors
- Remain accountable for editorial decisions and oversight
- Apply proportional, risk-based judgment when assessing AI use
- Intervene where accountability or transparency is unclear
- Escalate where integrity or policy boundaries are breached
Governance and Review
Existing research integrity, ethics, and editorial governance processes apply to all AI-related concerns. These policies will be reviewed periodically to ensure alignment with evolving technologies, regulations, and community expectations.