Introduction

At Articulate Marketing, we believe in the power of technology to drive innovation, deliver value to clients and make a positive impact on the world.

In line with our core values and our commitment to responsible business practices, we have developed (and continue to evolve) this AI Policy to address the risks and opportunities associated with generative AI.

This policy provides guidelines for clients and employees regarding the use of AI in our work.

Policy principles

• People first. We are human beings writing, designing, coding and consulting for human beings. We will ensure human oversight in all critical decision-making and creative processes involving AI.
• Transparency. Starting with this policy, we aim to be transparent and accountable about our use of AI. We will engage with clients to understand their specific requirements and concerns regarding AI.
• Responsible use. We aim to understand and mitigate the risk of bias, discrimination, hallucination and error that AI might create. We also respect privacy and data protection regulations in all our AI-driven processes.
• Non-reliance. We will not let our intellectual and creative muscles atrophy by over-reliance on AI. We will always bring our full human talent, experience, expertise and diligence to the work.
• Enthusiastic experimentation. We will invest in research, development and training to explore new applications and foster an environment where colleagues can experiment with AI technologies responsibly. We recognise that writing good prompts is, in itself, a creative act.

Risks and opportunities

Generative AI offers tremendous opportunities, but as our CEO Matthew Stibbe noted, it also comes with risks, and there are things it can’t (yet) do for us.

Opportunities

Tireless proofreading

Accelerated editing

Adjusting tone of voice

Drafting sales emails

Sales forecasting and analysis

Creative sidekick

Transcription and subtitling

Idea generation

Topic outlines

‘Explain to me like I’m 5’

Data analysis

Code snippets

Drafting workshop and interview frameworks

Placeholder copy for designers

Discovering best practices and new techniques

Optimising, debugging, adding accessibility to code

Making code more readable

Risks

Unconscious bias

Data protection and client confidentiality

Factual inaccuracy

Out-of-date information

Hallucinations

IP Infringement

Unreliability

Bland copywriting

Inconsistent imagery

Offensive imagery

Hidden text in imagery

Inaccurate imagery (humans with 6 fingers, etc)

Unreadable, buggy or hard-to-understand code

Accidentally adding a security hole in an app

What AI can’t do

EEAT - experience, expertise, authority and trustworthiness

Thought leadership

Storytelling and suspense

Deep understanding

Going beyond the prosaic

AI tools

Primarily, we use:

• ChatGPT (paid corporate licence so OpenAI doesn’t train on our data)
• Teams, Riverside and Loom (for transcriptions and summaries)
• Grammarly