The College recognises the importance and potential of all forms of AI systems, including generative AI, AI agents, and agentic AI, in the higher education sector.
These systems have exponential potential to increase productivity, foster innovation, and promote collaboration across a diverse range of domains. At the same time, the College acknowledges the immediate and future risks these systems pose to academic and research integrity, institutional reputation, research quality, data privacy and security, transparency, accountability, inclusion and equity, and; most importantly; the human essence of education.
In response, the College commits to responsible AI governance through clear policies and guidelines for staff, faculty, students, and the introduction of new AI systems. The College further commits to building institution-wide AI competency and to a continuous learning process, so that AI is governed and applied for growth and positive change rather than adopted uncritically.
The College has identified seven core principles in AI governance, forming the seven pillars of the AAEEHST framework that guide a safe and responsible approach to AI.
The College upholds academic and research integrity through clear, timely guidance on the responsible use of AI in academic and research work. AI-assisted work must be acknowledged and must not misrepresent authorship. Questions of authorship, ownership, and intellectual property in AI-generated work are addressed under this principle.
The College ensures that the AI systems it uses are accessible to users with diverse needs and abilities, and puts in place mechanisms to assess and reduce the bias and discrimination that can arise from these systems.
The College complies with relevant international, regional, national, and institutional ethical and legal standards, and consistently promotes the ethical and responsible use of AI. Consistent with GNH, the College is mindful of the environmental and energy footprint of the AI it adopts.
The College encourages the safe and responsible use of AI among its stakeholders for exploration, collaboration, and innovation, so that it remains responsive and adaptive to the rapidly evolving AI landscape.
The College maintains human oversight over all AI systems, with ultimate decision-making authority remaining with people. AI is used to augment and complement human judgement and expertise, not to replace it. Final authority in high-stakes matters; including admissions, assessment and grading, staff evaluation, financial, and disciplinary decisions; always rests with a responsible human.
The College continuously assesses the AI tools in use to protect the stability of interdependent systems, and maintains robust practices for data privacy and security. Sensitive, personal, or confidential institutional data must not be entered into public or unapproved AI tools.
The College maintains transparency about its adoption and application of AI and acknowledges its use in relevant contexts. Members of the community are expected to disclose material use of AI in their work. The College communicates its AI guidelines clearly to all stakeholders.
Certain uses of AI are prohibited for all members of the College community, in any context; academic, research, administrative, or personal use of College systems and data.
These red lines reflect risks that Bhutan’s national frameworks also single out, including misinformation, deepfakes, and discrimination. Breaches may be addressed under the College’s codes of conduct and disciplinary procedures.
Using AI to harass, bully, threaten, defame, impersonate, or demean any person.
Using AI in ways that discriminate against or unfairly disadvantage any individual or group, including allowing biased AI output to influence decisions.
Intentionally generating or spreading false, misleading, or deceptive content; whether about people, research findings, or the College.
Creating deepfakes or manipulated images, audio, or video of real people without their knowledge and consent.
Entering another person’s personal or sensitive data, or the College’s sensitive data into AI tools without authorisation, or using AI to profile individuals.
Using AI to produce content that is unlawful, that endangers safety, or that breaches national law or other College policies.