AI Policy
This statement explains Firebrand’s use of Artificial Intelligence (AI) for internal and client work.
We are excited about the potential of AI technologies to develop and deliver marketing programs more efficiently. Equally, we recognize that these technologies are new, potentially presenting ethical, privacy, confidentiality, and copyright risks to our clients.
AI Policy
Accordingly, here are our policy guidelines to govern our use of AI-powered software.
1. Human oversight – All work will be directed by specific individuals who are ultimately responsible for the work product. No deliverables will be produced entirely by AI tools, and all work will be reviewed by our team for accuracy and authenticity.
For example, while written content may be created using GenAI technologies such as ChatGPT, this will only be in input to the final deliverable.
2. Confidentiality – Firebrand will take appropriate measures to maintain the confidentiality of client information in accordance with industry standards and its agreements.
For example, confidential material will not be added to the input fields of LLMs where such material may be ingested into the model itself.
3. Disclosure – We will disclose to clients on request how we are using AI tools to deliver their campaigns. At no point will Firebrand claim work entirely created by an AI to be that of a member of the team.
For example, we may use a tool such as Fathom or Otter.AI to record and transcribe calls for note-taking and to record actions. Clients can request that this be done manually if they prefer.
4. Best practices – We will endeavor to use AI where appropriate to make our campaigns as efficient and effective as possible. As part of this, we will train our team and share findings in our research with clients and the community.
For example, Firebrand will post summaries of its research on its blog as part of its Firebrand Labs team which reviews AI tools and their application to marketing.
5. Ethics – Firebrand will make every reasonable effort to understand, correct, and avoid any bias and error that AI can introduce to its outputs.
For example, Firebrand will check facts introduced to copy by tools such as ChatGPT and will not engage in any ‘deepfake’ activities that GenAI can create.
Current uses of AI
For transparency, here are areas in which Firebrand currently uses GenAI, though this is constantly changing so may not be comprehensive at the time of reading.
1. Copywriting – tools such as ChatGPT and Grammarly provide a range of content creation capabilities. We may use these for proofreading, summarizing, and headline production among others. No copy will be delivered without human editing and review.
2. Image creation – Firebrand uses tools such as Canva, Midjourney, Adobe Firefly for concepts, storyboards and image generation. No images will be delivered without human editing and review.
3. Transcription – tools such as Fathom, Zoom, and Otter.ai all provide meeting transcription and summarization.
4. Operations – tools such as Hubspot and Asana include chatbots and other AI-powered features to increase efficiency. Since we routinely use these and many other marketing and collaboration tools, we will take advantage of these features.
5. Communication – to the extent that tools such as Microsoft 365 offer AI or co-pilot features to automate workflows, we will judiciously use these to accelerate communication.
Questions
If you have questions about our use of AI, please contact us at [email protected], and we will be happy to explain our use and approach further.