Artificial Intelligence (AI) Policy

In applying the six ethical and trustworthy AI principles, those covered by this policy must adhere to the following general guidelines:

1. Consideration: Assess the full impact of AI use, recognising potential biases. For example, students should carefully interpret and apply AI results.
2. Ethical Assessments: Follow the School’s Research Ethics and Governance Code, including obtaining Academic Board approval when necessary.
3. Human Intervention: Exercise due care to uphold academic integrity and comply with legal requirements.
4. Continuous Monitoring: Regularly monitor AI usage to identify and address unintended consequences, seeking ethics approval if needed.
5. General Duty of Disclosure: Promptly and properly disclose all AI use to the School, such as in bibliographies and through the AGS disclosure form. 

All AI usage must be disclosed to the School using appropriate processes. These processes can include methods such as references, bibliographies, or by completing an AI usage form on the AGS referred to below. The AGS operates as the School’s governance and assurance system-of-record for AI disclosures and registers. It is essential that every instance of AI usage is accurately recorded and communicated to ensure compliance with the AI principles and to provide auditable evidence of decisions. Additionally, the manner in which AI is used will determine the specific way in which it needs to be disclosed.

In an information governance context, accountability for AI-related information assets and datasets sits with designated Data Owners within each functional domain Data Owners are responsible for ensuring lawful basis, transparency, retention requirements, and appropriate safeguards. 

LSI adopts a consistent institutional approach to artificial intelligence in summative assessment. Taught modules assess students through complementary components that establish independent academic achievement and the responsible application of AI within professional and academic contexts, in accordance with the approved module specification and Assessment Regulations.

1. AI Prohibited — Closed-Book Examinations

All invigilated examinations are closed-book assessments in which AI use is prohibited. Students must demonstrate their own knowledge, understanding, critical analysis, reasoning and judgement independently, without using generative AI, AI-assisted tools, notes, textbooks, websites, external communications or other unauthorised resources.

Examinations will be conducted under LSI's approved examination security and invigilation arrangements, including a lockdown browser where required and technically supported. No Module Leader or other member of staff may authorise AI use in an examination through an assessment brief or informal instruction. Approved reasonable adjustments remain applicable.

2. AI Assistive — Coursework

All coursework assessments must require meaningful and appropriately defined engagement with AI, aligned with the approved module learning outcomes. Where AI is classified as Assistive, students must use authorised AI tools for specified supporting activities, such as challenging assumptions, identifying questions for further investigation or critically testing preliminary ideas.

AI must not replace the independent research, analysis, evaluation, reasoning or judgement that the coursework is designed to assess. The assessment brief must specify the required AI activity, permitted and prohibited uses, evidence of engagement and applicable disclosure requirements.

3. AI Integral — Coursework

Where the approved learning outcomes support assessment of relevant AI-enabled professional or academic skills, coursework may be classified as AI Integral. Students must actively direct, test, critically evaluate and, where appropriate, refine the outputs of authorised AI tools as part of completing the assessment.

The assessment brief must explain how AI engagement contributes to demonstrating the relevant learning outcomes and identify the required evidence, including substantive prompts, outputs, critical evaluation and refinements where applicable. Students remain responsible for independently verifying information, exercising judgement and justifying their conclusions. Reproducing AI-generated outputs without appropriate critical engagement does not, by itself, demonstrate achievement of the learning outcomes.

4. Assessment Design and Approval

For each coursework assessment, the Module Leader must select either AI Assistive or AI Integral, based on the approved learning outcomes, assessment purpose and academic level. The proposed classification, required AI activities, permitted tools, prohibited uses, evidence requirements and marking arrangements must be documented in the assessment brief and supporting assessment materials.

The Module Leader must consult the Director of Education on the proposed classification before finalising the assessment. The Director of Education will confirm that the assessment remains academically valid, sufficiently rigorous, inclusive and consistent with the approved module specification. All assessments remain subject to LSI's established internal moderation, external scrutiny and approval arrangements.

5. AI Use in the Final Master's Project

The 60-credit Final Master's Project (FP10) must require meaningful and appropriately defined engagement with AI across the project as a whole, aligned with its approved learning outcomes, research objectives and assessment requirements.

The Module Leader must determine whether AI use is Assistive or Integral, and identify the project activities in which such engagement is academically appropriate. AI use need not be demonstrated separately in every assessment component, provided the project as a whole includes substantive, assessable and appropriately evidenced AI engagement.

Students remain responsible for their own research design, methodological decisions, analysis, critical evaluation, findings and conclusions. AI must not replace the independent intellectual contribution or academic judgement required to demonstrate achievement at Level 7.

The assessment brief must specify the authorised AI activities, applicable restrictions, disclosure requirements and evidence students must provide. Where relevant, students must demonstrate how they have critically evaluated, verified and refined AI-generated outputs.

6. Disclosure, Academic Integrity and Data Protection

Students must disclose their AI engagement using the method specified in the assessment instructions and LSI's disclosure requirements. This may include identifying the tools used, purposes of use, substantive prompts and outputs, relevant refinements, and an explanation of how AI informed their work.

Students must not present AI-generated work as their own independent contribution or exceed the authorised scope of assistance. Students must comply with LSI's Academic Misconduct Regulations, Information Governance Policy, Data Protection Policy and applicable copyright requirements. Permission to use AI does not authorise the uploading or disclosure of confidential, personal, copyrighted or restricted material without an appropriate lawful basis and authorisation.

7. Academic Standards and Student Preparation

Across examination and coursework components, marks must reflect achievement of the approved learning outcomes, relevant subject knowledge, critical evaluation, independent reasoning and justified academic or professional decisions.

Module Leaders must ensure that the teaching content and formative learning activities provide appropriate opportunities to prepare for the independent performance required in closed-book examinations and the responsible AI engagement required in coursework.

LSI will communicate these assessment requirements clearly before assessments are released, ensuring consistency, fairness, accessibility and transparency.

This policy and its principles also apply to all weekly study undertaken by students. Students must:

• Apply the principles and guidelines of this policy in their work.
• Use appropriate citations and explain their use of AI.
• Interpret and analyse AI results critically before using them.

Students are encouraged to maintain a reflective journal on their use of AI in learning. This practice helps them articulate their experiences and enhances their understanding of the learning journey.

Additionally, students should seek feedback on their AI-assisted work. This will help them improve their outputs and understand how to effectively blend AI contributions with their own personal insights.

Students with questions about using AI should consult their programme or module leaders, Personal Academic Tutors (PATs), or the Director of Education.

Academic staff may incorporate AI into their teaching; however, its use must align fully with the principles and processes outlined in this policy, including transparency and fairness.

The School actively develops AI tools for learning and teaching. For instance, it utilises intuitive knowledge graphs and incorporates AI into all teaching activities. Each week, automation and AI systems guide students through personalised learning experiences. Students engage with pre-designed challenges and receive targeted feedback from a virtual mentor, helping them identify areas for improvement and enhance their skills.

Staff conducting research or applying new AI-driven learning and teaching tools must adhere to relevant School policies, including those on research ethics and the principles outlined in this document.

Additionally, staff should disclose their use of AI by submitting the AGS form mentioned below.

The School may use AI in admissions to support eligibility checks, document review, fraud or inconsistency indicators, triage, summarising evidence, recommendations and applicant communications; final admissions decisions must be made or confirmed by authorised staff with meaningful human oversight unless a solely automated decision is separately lawful and the required safeguards are applied.

The Student Success Team may utilise AI to provide support but must ensure compliance with the principles outlined in this policy and with the School’s Data Protection Policy. Where AI uses student information, only the minimum necessary data should be processed, outputs should be reviewed by a human decision-maker, and personal or sensitive information should be anonymised or aggregated wherever possible. They should first consider the disclosure rules below. AI can be used to recommend tailored support for students based on their unique learning needs, predict performance issues to enable proactive interventions, and evaluate feedback to identify areas for improvement. By employing AI thoughtfully, the team can enhance the effectiveness of their support services and foster a more positive learning environment for students. They should ensure that any use of AI is properly disclosed by submitting the AGS form mentioned below. 

Staff and students must always disclose all AI usage in learning, research, and assessments by providing proper citations and references. They may also need to submit an AI Usage Form on the AGS. The principle is that any significant implementation of AI impacting teaching, learning, assessment, administration, or student support requires the completion of this form. This includes situations such as adopting new AI tools, reporting data breaches or misuse, developing courses that incorporate AI, implementing AI-driven assessments, or using AI for student support services. Any suspected or confirmed security incident or personal data breach involving AI systems must also be reported immediately through the School’s governed incident and breach reporting process in accordance with the Data Protection Policy

This form is separate to an ethics application and supports an AI Usage Register on the AGS, which records the individual using AI, the purpose of its use, specific tools employed, and adherence to policy principles. Maintaining the AI Usage Register provides an audit trail and data, supports reporting and continuous improvement, and facilitates annual reviews, thereby upholding the integrity and transparency of AI usage. AI disclosure records, forms, and supporting evidence are institutional records and are retained and disposed of in accordance with the School’s Data Retention Schedule and Policy.

Where an AI system or tool processes personal data, special category data, or information that may significantly affect individuals, the School will assess risks before deployment. This may include completion of a Data Protection Impact Assessment or equivalent risk assessment in accordance with the School’s Data Protection Policy. The assessment will consider fairness, bias, transparency, privacy risks, security controls, and proportionality. AI systems must not be deployed until identified risks are mitigated or formally accepted through the School’s governance processes.

The AI Policy is publicly available on the School's website and AGS.

The School will provide training for all staff and students on this AI policy and its principles, including:

1. Explainability: Guidelines on when and how AI should be used.
2. Transparency: Proper use of AI in academic work, particularly in summative assessments and citation.
3. Accountability: Compliance with relevant laws, including data protection and copyright.
4. Privacy: Understanding the implications of rights related to AI use.
5. Fairness: Ethical use of AI, including ethics assessments and supervisor responsibilities.
6. Robustness: Correct interpretation and application of AI results, addressing issues such as hallucinations and bias.

The School will continuously review this AI policy, including its guiding principles above and processes, as part of its annual evaluations to ensure it aligns with legal, ethical, professional, academic, and fair use standards for AI and other technologies.

The Executive Committee (EC) will conduct quarterly reviews and update the policy as necessary to reflect technological advancements and changes in the law.