Last update : 23 June 2026
At GlobetrottingFundraiser (GtF), artificial intelligence is treated as a working tool. Used well, it expands what one consultancy can deliver. It speeds up the repetitive parts of the work, sharpens early drafts, and frees up time for the thinking that moves legacy giving forward. None of that changes the standard the work is held to.
Legacy giving is built on trust. While we are glad to use AI where it genuinely helps, we use it with a clear set of limits, and this policy sets out exactly what those are.
This policy is written around the four pillars that guide everything at GlobetrottingFundraiser: Evidence, Trust, Capacity, and Integrity. If a use of AI cannot sit comfortably inside all four, it does not happen.
What is meant by AI
When this policy refers to AI, it means computer systems that simulate human cognitive tasks such as language generation, summarising, reasoning, translation, and pattern analysis. In practice, that means tools such as large language models and neural machine translation.
We work with two AI tools, and only two. The first is Claude, made by Anthropic, which is used for writing and analysis support. The second is DeepL, which is used for translation. Think of a tool like Claude as a very capable research assistant. It is fast, it is well read, and it is occasionally confidently wrong. That last part is exactly why a qualified human stays in charge of every output. The judgment, the sector knowledge, and the final word remain under the consultant’s purview.
How AI is used
AI may assist with the following tasks:
Data analysis and strategy support, such as spotting patterns, organising information, and pressure-testing ideas.
We do not currently use AI to generate images. Any images and photographs used are either our own, or licensed stock created by professional photographers. If AI is ever used to create an image, it will be clearly identified as AI-generated. All visual and design choices for GlobetrottingFundraiser are made by humans.
GtF’s toolset is deliberately limited to Claude and DeepL. We do not switch on or rely on the AI features built into the other software used to run the business, including Google, HubSpot, Canva, and Microsoft. Keeping the toolset small keeps it easy to govern, easy to disclose, and easy to trust.
Our principles, by pillar
Evidence. AI output is verified, not trusted on sight. Facts, figures, statistics, and citations are checked against credible sources before they go anywhere near a client deliverable. AI does not get to invent a statistic and have it slip through.
Trust. A human directs and reviews every output. AI never has the final word, and it never speaks to a client or donor on its own. Every prompt is written by me, every decision is made by me, and every piece of work carries my professional accountability.
Capacity. AI is a capacity multiplier, which is precisely what a specialist consultancy and the smaller charities I work with need. It helps level a field that has long favoured organisations with the biggest budgets. That benefit only holds if quality holds, so AI work stays at the draft stage and never ships as a finished product without human shaping.
Integrity. Confidentiality and fairness are non-negotiable. I do not enter confidential client information or personally identifiable donor data into AI tools. I direct AI away from bias and stereotype, and I review output with diversity, equity, inclusion, and belonging front of mind. Where AI cannot be used responsibly, it is not used at all.
What we will not do
Privacy and compliance
GlobetrottingFundraiser operates from Quebec and works with organisations across Canada and beyond. AI use here is handled in line with Canadian privacy law, including the federal Personal Information Protection and Electronic Documents Act (PIPEDA) and Quebec's Law 25, which governs the protection of personal information in the private sector. Where a client operates under another jurisdiction, such as the UK or EU, we take their applicable requirements into account as well.
The simplest safeguard is the one we rely on most: sensitive personal and donor information does not go into AI tools in the first place.
Keeping this current
AI is moving quickly, and a policy written today will need revisiting. We review this policy regularly and update it as the tools, the risks, and the sector's expectations evolve.