AI Tool Guidance

Guidelines for the Use of Generative AI Tools at Tuck

Tuck School of Business: Guidelines for the Use of Generative AI Tools 

At Tuck, we are committed to fostering innovation while upholding the highest standards of academic integrity, data privacy, and ethical responsibility. The following guidelines are designed to help our community navigate the evolving landscape of generative AI tools responsibly and effectively. These AI guidelines should be utilized in conjunction with the wider Dartmouth Information Technology Policy. 

1. Safeguard Confidential, Sensitive Information, and Privacy 

Protecting confidential and sensitive information is paramount. This includes data such as student records, financial information, proprietary research, and personal identifiers. In accordance with Dartmouth’s Information Security Policy, all members of the Tuck community are responsible for ensuring confidentiality, integrity, and availability of institutional data. 

When using generative AI tools, adhere to the following guidelines: 

  • Avoid Inputting Sensitive Data: Do not input confidential or sensitive information into publicly available generative AI tools, as these platforms may not guarantee data privacy and could expose information to unauthorized parties.
  • Use Approved Tools: Only utilize AI tools that have been vetted and approved by Tuck IT and Dartmouth’s Information, Technology & Consulting (ITC) department for handling sensitive information. Approved tools are designed to comply with Dartmouth’s data protection standards. Please refer to Tucks   AI Tools and Resource Hub to see a list of current tools.
  • Understand Data Classification: Familiarize yourself with Dartmouth’s data classification levels (DISC 0 to DISC 3) to determine the sensitivity of the data you handle. Ensure that the AI tools you use are appropriate for the classification level of the data involved. Other Dartmouth policies around information security, privacy, intellectual property, data governance and more can be found here.
  • Consult Tuck IT for Guidance: If you are uncertain about the appropriateness of an AI tool for specific data types, consult with Tuck IT. We can provide guidance on data protection requirements and assist in evaluating AI tools for compliance with institutional policies. Visit: https://services.tuck.dartmouth.edu to get in touch.
  • Tuck Leadership does not monitor, review, or have visibility into individual prompts, conversations, or content entered into approved enterprise AI tools. Usage insights may be used to understand overall consumption trends and service sustainability, consistent with industry norms for enterprise technology management. 

2. Verify AI-Generated Content Before Use 

AI-generated outputs can contain inaccuracies, biases, or fabricated information. It is your responsibility to critically evaluate and verify the accuracy and appropriateness of any AI-generated content before sharing or incorporating it into your work. 

To ensure the reliability of AI-generated content: 

  • Cross-Check Information: Always corroborate AI-generated information with credible sources, especially when dealing with factual data, statistics, or references.
  • Assess for Bias and Sensitivity: Be vigilant for potential biases or culturally insensitive material in AI outputs. AI models may inadvertently produce content that reflects societal biases present in their training data.
  • Maintain Transparency: If AI tools are used in your work, disclose their usage appropriately, ensuring transparency in your academic or professional contributions. 

The use of AI-generated content requires careful human oversight. You must verify and, if needed, revise content to maintain accuracy, appropriateness, and alignment with institutional values before dissemination. 

3. Adhere to Academic Integrity Standards 

  • Familiarize yourself with Tuck’s Academic Honor Principle, which emphasizes integrity and honesty in all academic endeavors. Refer to the student handbook for more detailed information.
  • Faculty should clearly communicate their expectations regarding the use of AI tools in coursework and assignments.
  • Students should seek clarification from faculty members if uncertain about permissible uses of AI tools in their academic work.
  • Guiding Principles on the Use of AI Tools in Academics:
    • Faculty should strive to be clear in their syllabi and elsewhere about where AI tools are allowed and where they are not allowed in their courses, including their possible use in supporting activities such as brainstorming and proofreading. This may also include restrictions on the use of course materials for training or as other input for generative AI systems.
    • Students should take the time to understand the AI-related policies in each of their courses and seek input from relevant faculty if there is any perceived ambiguity. Tuck’s Honor Principle governs adherence to course policies, including those related to generative AI.
    • Students should be transparent about any use of AI tools in their coursework. As with other source material, use of AI tools should be cited appropriately. 

4. Remain Vigilant Against AI-Enhanced Phishing and Security Threats 

  • Be aware that AI technologies can be used to create sophisticated phishing attempts and other security threats.
  • Exercise caution with unsolicited communications and report any suspicious messages to Tuck’s IT security team. 

5. Utilize Approved AI Tools for Tuck Activities 

Ensure that the AI tools you employ for academic and administrative tasks meet institutional standards. 

  • Use Vetted Tools: Employ only those generative AI tools that have been reviewed and approved by Tuck’s IT department for official academic and administrative tasks. List of vetted tools can be found on the AI Tools and Resources Hub.
  • Seek Approval for New Tools: Consult with Tuck IT before adopting new AI tools to ensure they meet institutional standards and compliance requirements. Reach out to Tuck IT: Tuck IT Service Portal 

6. Understanding AI Tool Consumption and Availability 

Generative AI tools are evolving rapidly, and not all tools are offered under the same access or cost models. Some enterprise‑supported AI tools operate on a consumption‑based model, meaning their availability and features are tied to overall usage levels, licensing structures, or vendor‑controlled releases. 

Tuck IT, in partnership with Tuck senior leadership, manages access to these tools to ensure equitable use and to keep institutional spending within established budgetary constraints. This management is based on aggregate usage information provided by vendors, not on individual user activity. 

As a result: 

  • Tool features and model availability may change over time. In some cases, new capabilities may be introduced gradually, limited, or unavailable due to vendor decisions that are outside of Tuck IT’s direct control.
  • Access levels may vary by tool. Some AI platforms may offer different tiers, models, or capabilities depending on consumption limits or licensing terms.
  • Temporary adjustments may be necessary. To support equitable and sustainable access for the broader Tuck community, Tuck IT may occasionally adjust availability, pause specific features, or prioritize certain use cases. 

Currently, the Tuck community has full, unlimited access to models and features within our Enterprise Microsoft 365 Copilot platform. Other enterprise AI tools may operate differently and are subject to the considerations outlined above. For a detailed breakdown, see AI Tools and Consumption at Tuck. 

Tuck IT is committed to transparency and clear communication as changes occur. Our goal is to balance innovation, fiscal responsibility, privacy, and broad tool access while supporting the academic and administrative needs of the Tuck community. 

7. AI Coding and Software Development Tools 

Some generative AI tools can write, modify, or execute code. Because these tools can directly affect systems, data, and security, their use carries additional considerations beyond general AI assistance. 

Use of AI tools for coding, scripting, software development, or system automation is governed by separate, role‑specific guidance developed by Tuck IT. That guidance outlines approved tools, appropriate use by persona (students, faculty, staff, and IT professionals), and required approval paths for advanced or autonomous capabilities. 

Community members engaging in AI‑assisted coding or development work should review and follow the AI Coding Guidance and Guardrails before using these tools for institutional work. 

This document focuses on general principles for responsible AI use and does not replace more specific technical standards or guardrails. 

8. Exercise Caution with AI-Powered Meeting Assistants 

  • Obtain Approval and Consent: Do not use AI-powered meeting assistants (e.g., transcription or summarization tools) during Tuck meetings unless the tool is approved, and all participants have provided informed consent.
  • Zoom AI Companion & Microsoft Teams (Transcription and Copilot in Teams) are approved tools.
  • Avoid Sensitive Discussions: Avoid employing such tools in meetings that involve confidential or sensitive information.
  • Review and Manage Outputs: Review and, if necessary, correct AI-generated summaries before distribution, and delete them when they are no longer needed. When in Doubt, Get in Touch! 

9. Use Caution with Autonomous AI Agents and Automated Workflows 

Some emerging AI tools allow users to create automated agents that can perform tasks on their behalf, such as retrieving information, interacting with applications, or automating workflows. 

While these tools can offer productivity benefits, they may also introduce additional risks related to data access, security, and system integrity. 

When using AI agents or automation tools: 

  • Do not connect AI agents to institutional systems, data repositories, or communication platforms without review and approval from Tuck IT.
  • Do not grant automated AI systems access to confidential or sensitive institutional data unless the tool has been approved for that purpose.
  • Maintain human oversight. Users remain responsible for any actions taken by AI systems operating on their behalf. 

If you are considering deploying AI-driven automation or agents for institutional work, consult Tuck IT to ensure alignment with Dartmouth policies and security practices. 

For questions, further guidance, or to discuss the use of generative AI tools at Tuck, please reach out to Tuck IT: 

Tuck IT is here to assist you in navigating AI tools responsibly and in alignment with Tuck and Dartmouth’s policies.