AI policy

This document has been published to identify how Keystone Marketing Limited uses AI applications ethically, legally, and effectively so that we protect our clients and always demonstrate best practices. AI is a rapidly advancing technology, and we will update this policy regularly.

We are committed to identifying and capturing opportunities arising from emerging technologies. We are encouraged to be curious about new technology, expand our understanding of how AI can be used and how it works, and use it within the parameters set out in this policy.

While we know the potential offered by AI, we also acknowledge the limitations and pitfalls and continually seek to uphold factual accuracy, observe ethical usage, and continuously refine our approach to using it.

Our use of AI will support three core principles:

Quality: We’re cautious and circumspect in our approach to working with generative AI, using it for research and ideation. We will check and cite with care, to avoid being susceptible to bias and misinformation.
Protection: We will follow developments in legal regulation closely to ensure we are protecting our content and clients from copyright issues and fair use of data.
Privacy: Keystone will never input sensitive information or private client data into generative AI. We will adhere to the principles of GDPR, being transparent in our use of generative AI and how we collect and store data.

ESSENTIAL TERMS

Keystone Marketing Limited (from here, called Keystone) is a marketing consultancy based in St Neots, Cambridgeshire, UK. We develop marketing strategies, run marketing campaigns, create content such as copywriting, social media posts, graphics, video and podcasts, and deliver training in person and online.

Artificial intelligence, or AI, refers to the simulation of human intelligence in machines that are programmed to think and act like humans.

Generative AI is a broad label for any type of artificial intelligence (AI) that can be used to create new text, images, video, audio, or code.

Large Language Models (LLMs) are part of this category of AI. They are trained on vast amounts of existing data and generate human-like text.

ChatGPT, Claude and Google Gemini are publicly available web-based versions of generative AI that allow users to enter text prompts and seek a view from the system or to ask the system to create textual output based on a given subject. They allow individuals to summarise long articles, get an answer of a specific length to a question, or have code written for a described function.

BEHIND THE SCENES

We always share with each other how we’re using AI in our daily work. We want to use AI to support our creativity, whether that’s AI embedded in the platforms we currently use or a Keystone-approved* list of specifically developed AI applications. We believe that people demonstrate integrity, understanding and empathy, and we want to use AI to support that, not replace it.

*Approved applications are:

ChatGPT 4o Team
Otter.ai
DALL.E

We support sustainable practices in digital activity.

Training AI on large data sets, using them to capture information, and then storing this data has a large carbon footprint in itself. We will carefully consider our usage, so we strike a balance between realising the incredible potential of AI and lessening the environmental impact where possible.

USING AI TEXT GENERATORS

Generative AI like ChatGPT can process vast amounts of information to swiftly summarise complex topics.

When using text generators to research a topic, Keystone will always fact-check and cross-check the resulting material to ensure false information is not mistaken as fact. We will never put sensitive and private client data into a text generator.

We may use our Keystone-approved text generator to spark ideas and inspiration.

AI can support us in project research and the ideas generation process, similar to how we ask a colleague for an idea on how to approach a creative task.

We will always prioritise quality and authenticity in the content we produce.

We will be critical and responsible users of generative AI – and that means not publishing anything written completely by them and being aware of the many reasons why we shouldn’t. Material gathered from a text generator will not be copied and pasted into either our internal or public-facing material without human editorial processing.

Through human editorial processing, we will always:

Ensure all content is factually correct and ethically sound
Edit wording to the appropriate tone of voice and brand guidelines
Check for biases and stereotypes
Avoid plagiarism

With this in mind, we will not publish any written content that is 100% produced by generative AI. We will always apply brand guidelines, fact-check responses, and review and refine in our own words.

USING AI AUDIOVISUAL TOOLS

Keystone may use image tools to correct and make minor edits.

Image editing with Keystone-approved AI can speed up our work. We will use tools like DALL.E to make minor changes to a photo, to make it more usable without changing the subject matter or original essence.

We will use image editing AI ethically.

We do not use any image or video editing tools to change the essence of any original visual content, on our website, social media, or anywhere else. For example, we do not change the expressions, appearance, ethnicity, or any other core features of those captured in photography, video and images.

Keystone may use audio and video tools to remove distractions for audiences.

Just like image tools and text generators, AI audio and video applications can also help us save time with creative projects. For example, we can save hours of editing time by using in app audio clean-up tools to remove unwanted background noise or using video clean-up apps to remove distractions in shots in seconds.

AI may be used for creating subtitles in video editing programs.

This saves us time, enabling us to pass the cost savings onto our clients. Some social media platforms will use AI to create voiceovers to go with the subtitles too. We will always proofread the final content that goes out to ensure accuracy.

Keystone will never use voice clone generators or create deepfake videos.

We do not use AI to recreate the voice or appearance of anyone, as there are major ethical implications and legal risks attached to this.

RESEARCH, ANALYSIS AND IDEATION

Keystone does not input any sensitive, private or embargoed information into AI

Keystone does not input sensitive, private, or embargoed information into AI. While text generators like ChatGPT allow users to input articles or data to generate insights or prompts—such as creating a social media post based on an article or analysing data for trends—Keystone remains committed to safeguarding privacy and intellectual property.

With ChatGPT 4o, inputs are not used to train the AI model, providing an added layer of security. However, Keystone will continue to exercise caution by only inputting information that is already in the public domain. We will not input confidential or restricted data, adhering to the same principles we follow for external emails, public discussions, or social media sharing.

USING AI TO RECORD MEETINGS

We will not record or store content without consent.

Our AI transcription tool Otter.ai, enables meetings to be transcribed and the audio recorded and stored.

We may request consent to record an online meeting using Otter.ai or Google’s Gemini when we need to gather information for content creation, for example an interview to help us write a blog post, to transcribe client meetings, or other business discussions for internal use. We will give advance notice and mention this at the start and the end of the call, including an explanation of where the information will be stored.

MANAGING RISK

To manage risks effectively, we focus on several key areas:

Ensuring content quality to prevent ineffective copy and visuals that could hinder marketing objectives across channels such as organic search and email marketing.
Safeguarding our reputation by avoiding poor-quality or off-brand content.
Protecting customer privacy with strict data management practices.
Addressing ethical risks by promoting diversity and Inclusion and avoiding any form of discrimination.
Mitigating intellectual property risks by securing legal and proprietary information.

Our approach ensures that all AI-generated materials align with our client’s marketing goals, uphold their brand’s integrity, and comply with legal and ethical standards.

Policy created: 06/01/25

Date of last policy review: 06/01/25