Dartmouth College Provost's Undisclosed AI-Assisted Writing Was Found to Have Breached Publication Policies

A faculty review at Dartmouth College in Hanover, New Hampshire, found that provost Santiago Schnell used ChatGPT and Claude in published writing without required disclosure and fell short of institutional standards. The review examined AI logs and correspondence rather than detector scores. Schnell resigned and was leaving the college after the findings were released.

Risk classification

• Primary risk domain: 4 Malicious actors
• Primary risk subdomain: 4.3 Fraud, scams, and targeted manipulation

The provost engaged in AI-facilitated academic cheating by using Claude and ChatGPT to edit papers and write op-eds without disclosing the AI assistance, violating journal and newspaper standards for personal/professional advantage.

Causal factors

• Entity: Human
• Intent: Intentional
• Timing: Post-deployment

The incident was caused by a human decision (by Santiago Schnell) to intentionally use AI tools in his writing and intentionally omit the required disclosures.

EU AI Act risk tier

• Risk tier: 3 Limited Risk

Limited Risk: The incident involves the use of chatbots (Claude and ChatGPT) to generate and edit text and figures, which falls under Risk Level 3 due to transparency obligations requiring disclosure when AI-generated content is used.

AI system and alleged parties

• AI system: ChatGPT, Claude (Anthropic, OpenAI)
• AI purpose: Writing Assistant; Code Generation
• Behaviour type: Assistant
• Alleged developer: OpenAI, Large language model developers, Anthropic
• Alleged deployer: Santiago Schnell, Chatbot users, Academics
• Alleged harmed parties: Universities, Santiago Schnell, Educational communities, Dartmouth College, Academic integrity

Harm severity

Severity ratings are based on this scale.

Highest direct severity in any category: Negligible. Severity is scored from Negligible to Catastrophic in each harm category, for harm the reports describe as caused directly or indirectly by the AI system.

• Physical: direct Negligible, indirect Negligible
• Infrastructure: direct Negligible, indirect Negligible
• Property: direct Negligible, indirect Negligible
• Financial: direct Negligible, indirect Negligible
• Environmental: direct Negligible, indirect Negligible
• Malicious content: direct Negligible, indirect Negligible
• Differential treatment: direct Negligible, indirect Negligible
• Civil rights: direct Negligible, indirect Negligible
• Democracy: direct Negligible, indirect Negligible
• Privacy: direct Negligible, indirect Negligible
• Psychological: direct Negligible, indirect Negligible
• Epistemic: direct Negligible, indirect Negligible
• Child sexual exploitation and abuse: direct Negligible, indirect Negligible

People affected

• Occurrences reported: 1
• People reportedly harmed: 1
• People reportedly exposed: 1

Potential causes

Management

• Inadequate Executive Oversight: University lacked proactive oversight of executive AI use.
• Failure to Model Ethical AI Use: As Provost, the author failed to set a transparent AI example.

Technology

• Unmonitored Enterprise AI Access: Easy access to Claude and ChatGPT enabled undisclosed assistance.
• Lack of Automated AI Detection: Publishers lacked automated tools to detect AI-generated content.

Data Inputs

• AI Prompts for Academic Editing: Prompts asked Claude to find scientific issues and troubleshoot code.
• AI-Generated Figures and Code: Manuscript text and code were input to generate figures and edits.

Human Factors

• Undisclosed AI Assistance: The author chose not to disclose AI usage to publishers.
• Misjudgment of Ethical AI Use: Failing to recognize that AI refinement violated integrity codes.
• Evasive Response to AI Inquiry: The author delayed and failed to answer AI-related audit questions.

Process and Methods

• Inadequate AI Policy Workflows: Absence of clear workflows for verifying AI disclosure in submissions.
• Unregulated AI Editing Methods: Substantial AI editing occurred without formal tracking or logs.

Regulatory Environment

• Journal AI Disclosure Rules: ChemSystemsChem mandated AI disclosure which the author bypassed.
• Newspaper AI Limitations: The Washington Post standards prohibited undisclosed AI refinement.

Information quality

• Classification confidence: High
• Reason for confidence: The report clearly details the specific AI tools used (Claude and ChatGPT), the context of their use (academic and public writing), the specific violations (failure to disclose), and the direct consequence (resignation of the provost). There is little ambiguity regarding the facts of the incident.

The resignation of the Dartmouth College provost over undisclosed AI use in academic and public writing is an ethical and academic integrity issue with negligible national security implications.

• Overall national security impact: Negligible
• Scope: Single nation
• Primary target: United States
• Alleged perpetrator: Santiago Schnell

Threat characteristics

• Imminence: Long-term. The incident has already been resolved through the resignation of the provost, posing no ongoing or immediate national security crisis.
• Autonomy: Human-controlled. The AI tools were used purely as writing and coding assistants, with the human user retaining full control over the final submissions.
• Novelty: Established threat. Undisclosed use of AI in academic and journalistic writing is an established ethical issue rather than a novel national security threat.

Impact by dimension

• Physical security: Negligible. No physical systems, infrastructure, or human safety elements were involved or threatened in this academic integrity incident.
• Information security: Negligible. The incident involves personal academic misconduct and does not constitute a coordinated disinformation campaign or intelligence compromise.
• Sovereignty: Negligible. The event occurred within a private academic institution and has no impact on state authority, government operations, or constitutional processes.
• Economic security: Negligible. No strategic technologies were stolen, and there was no impact on national financial systems or economic security.
• Societal stability: Negligible. The incident is limited to an individual academic integrity violation and does not threaten social cohesion, civil liberties, or public safety.