Why do I need an AI ethics policy?
Figure skating used to score two things: did you get it right? And did it move anyone?
I use AI for work every day, and I help clients implement it successfully. This document says plainly what that means: where AI plays a role, where it doesn’t, and the commitments you can expect from me.
In practice, my AI use falls into three categories:
1 — AI-generated drafts: I tell you up front
Where AI produces a first draft that forms the structural basis of your work, a positioning document, communications plan or set of copy, I make that clear. These documents are a starting point and I review and edit every draft, taking full professional responsibility for what you receive.
2 — AI-assisted research and synthesis: ask and I’ll tell you
Where AI researches, summarises or surfaces material that then feeds my own analysis, I treat it like any professional research tool — a database, a search engine. Ask and you’ll get a full answer.
3 — AI as quality check: no disclosure
Spelling, grammar and formatting checks with no role in shaping the content. Equivalent to a spellchecker.
I am always the decision-maker. AI may support the work; but I always make the final call.
I do not use AI in a way that compromises client confidentiality, creates unagreed personal-data risk, or removes human accountability from consequential communications.
I treat personal information and identifiable material with care, and do not use AI in ways that create unagreed risk to a person’s privacy or reputation.
Bias and representation. AI can reproduce patterns and stereotypes from its training data. I do not treat its output as neutral. When work affects how people, audiences or markets are represented, I check it for assumptions and bias before it is used.
Accuracy and evidence. AI-generated output is not evidence in itself. I check material facts, sources and claims before relying on them in client work.
Environmental impact. I use AI purposefully, not by default. I avoid unnecessary generation and repetition, and choose a simpler approach where it will achieve the same result. I keep this under review as the evidence and tools improve.
Strategic recommendations. AI can generate options, surface research and structure analysis. The advice itself is mine based on decades of experience.
Brand positioning and messaging. A tool can create a very rough draft. The judgement that commits your brand to a direction carries my name, my creativity and my accountability.
Sensitive communications. Crisis work, reputation management and anything concerning a named individual — I draft, review and sign off personally.
Pricing, proposals and commercial commitments. My commercial terms get full human review and sign-off, every time.
Ghostwriting without the subject’s knowledge. I am happy to draft content in your voice, provided you are involved in the process. I will not publish work under someone’s name unless they are fully aware of how it was created.
I use AI both to accelerate work and to generate better inputs for strategic judgement. First drafts arrive sooner, which buys more time for the strategy underneath them. For selected questions, specialist AI workflows help me with tasks like pressure-testing positioning, exploring how different B2B buyers may respond and surfacing patterns worth investigating.
They do not replace judgement; they give it more to work with.
The angle, the risk judgement, the read of your situation and the sense of what a piece of work is actually for, is just as important a part of my job as it’s always been.
That is the standard I bring to client implementation work too: AI should reduce low-value effort and create more room for judgement, relationships and higher-value work.
Item | Detail |
|---|---|
AI ethics owner | Emma Gordon, Director, Bydand Strategic Communications Limited |
Raising a concern | Email emma@bydandstrategic.com or raise it with me directly. I will explain, plainly, what role AI played in a piece of work and how I reviewed it. |
Companion document | My AI usage policy — the operational rules on approved tools, data handling and sign-off |
I review this quarterly — October 2026, January 2027, April 2027, ongoing — and update it whenever my AI practice materially changes, a new disclosure obligation arises, or a client raises a concern that shows it needs sharpening. The current version is always at bydandstrategic.com.
Figure skating used to score two things: did you get it right? And did it move anyone?
The prompting guidance changed. Did you notice?
Most marketing professionals know AI matters. The problem isn't ability or tools, it's knowing where to begin. This article looks at why 75% of organisations still have no AI roadmap, and introduces a Google Gem built to help you find your starting point.