AI Usage Policy

The PEM Board has formally adopted the PEM AI Usage policy to govern our use of AI. This policy is a public document and will be regularly updated to remain aligned with the fast-changing AI landscape.
At PEM, we use Artificial Intelligence (AI) as a supportive tool to enhance efficiency, clarity, and insight in our work. AI is never used as a substitute for professional judgement, accountability, or human expertise.

1. How We Use AI

AI may be used to:

  • Supporting language review, including grammar and clarity improvements.
  • Creating, editing, or processing images, audio, and video materials for internal use, where rights and licenses are clear.
  • Supporting the analysis of large datasets to identify patterns, trends, and potential causal relationships across investments and outcomes, subject to appropriate data-protection controls.
  • Assisting with comparative and predictive analyses for internal benchmarking purposes.

AI is not used to:

  • Generate deliverables that are presented to clients as original PEM analysis without human authorship, review or adaptation.
  • Make autonomous decisions that have legal, financial or ethical implications for clients or staff.
  • Create deceptive, defamatory, discriminatory or misleading content, including deepfakes or content “in the style of” identifiable individuals.
  • Circumvent contractual, copyright or licensing restrictions.

2. Human Oversight and Accountability

  • Humans remain fully responsible for all work delivered.
  • No deliverables are produced or issued by AI.
  • AI outputs are treated as aids, not final products.

All AI-assisted outputs are reviewed, validated and finalised by partners and consultants. Where AI-generated content is used as an input (e.g., as a draft or source of inspiration), partners and consultants remain responsible for ensuring that the final product is accurate, original and appropriately referenced.

3. Transparency

PEM is committed to transparency regarding our use of AI. If applicable and requested by the client, an AI statement explaining how and for what purpose AI was used will be provided but is not formulated by default.

4. Data Privacy

  • At PEM, we share data with an AI tool only when it is in the public domain. When the data is not in the public domain, we apply a precautionary principle.
  • At PEM, we are committed of ensuring ethical AI use and protecting client data.

5. Principles for Responsible AI Use at PEM:

  • Lawfulness, fairness and non-discrimination. AI use must avoid discriminatory outcomes.
  • Transparency and explainability. If applicable and requested by the client, an AI statement explaining how and for what purpose AI was used will be provided but is not formulated by default.
  • Data security. Appropriate organisational measures are applied to protect it.
  • Accountability and human judgement. Humans remain responsible for all work delivered.
  • No impersonation or deceptive use. AI must not be used to generate content that impersonates individuals or organisations, or that misleads clients or the public.

6. Review

The PEM AI Usage Policy will be reviewed and updated every six months to ensure that it remains current and effective.
Last updated: March 2026.

7. Acknowledgement and Compliance

The Managing Director and the Chairperson of PEM agree that PEM will comply with the AI Usage Policy.

Newsletter