By Monica Davidson

In the last few months, the conversation for many of us has moved from “Are you using AI” to “HOW are you using AI?” As creatives, we’re both angry at the IP thievery and environmental damage that our new robot collaborators are doing, and at the same time (perhaps shamefully) using AI agents to help us draft difficult emails, double-check our spelling, or enhance our photos on social media.

If you are using AI in your creative business, and if you have clients or customers that you want to build trusting relationships with, then AI Protocols are a must. Yes, this sounds stuffy and corporate, but as generative AI floods our work world, ignoring it isn’t a strategy anymore. People want to know how we are using AI, and we can tell them in a way that reinforces our authenticity and our ethics. 

AI Protocols don’t need to be a 50-page manual. They can be a simple set of boundaries that protect your creative integrity, reassure your clients, and keep your sanity intact. They are also increasingly becoming a part of what government clients need for transparent procurement, which means there will probably be a flow-on effect to smaller government organisations and corporate entities.

What Clients Want from AI Protocols

From my experience, it seems clients and customers are starting to ask questions about the way we use AI:

• Which AI tools are you using, and what do you use them for?
• Why are you using AI instead of a human for those tasks?
• What guardrails are in place to protect creativity and confidentiality? 
• How are you protecting legal rights and overcoming the inherent biases in generative AI?
• How do you ensure that human judgement ultimately guides AI work? 

If your AI protocols go even a small way towards answering those questions, you’re making a good start. 

Remaining Ethical in Uncertain Times

Your AI protocol can be an opportunity to share your values and –  if this is your opinion – assert your firm stance against non-consensual data scraping, copyright exploitation and the environmental impact of large data centres. You can also ensure that you have an internal policy, shared in your protocols, for checking cultural biases and representational harm (ask Chat GPT to create an image of a ‘boss’ if you want to see this in action).

Ultimately, AI protocols are also a good opportunity to use genuine transparency as an opportunity for generosity with your clients and customers. Clients don’t necessarily hate AI, they hate feeling tricked. Reassure them by including your AI protocols in your quotes and contracts, even pop them on your website if you think it’s warranted. 

A Simple Example of AI protocols. 

Here is a very short and simple example of AI protocols for a fictional video production business. And yes, I did use Google Gemini to help me craft these words. 

• Human-Led Storytelling: AI is our production assistant, not our Director. We use AI for rough shot-list ideation and mood-board inspiration, but human eyes craft every script, edit every frame, and direct every scene.
• Training Toggled OFF: All team AI accounts have data-training turned off. Unreleased video footage, interview transcripts, and confidential client briefs are never used to train public AI models.
• Ethical Production: We stand firmly against voice-cloning without consent, deepfakes, and generative video tools trained on stolen independent art. We also ensure proper representation of diversity and actively work to correct cultural biases and representational harm in our AI agents. 
• 30-Second PM Tagging: If AI was used to help brainstorm a concept, we drop a 30-second note in Asana detailing the tool, the prompt, and how the human team refined the final idea.

At the end of the day, AI is likely here to stay. If you are using it in any way, whether that’s to organise your weekly calendar or amplify your creative ingenuity, AI protocols are a must. They will not only reassure your customers and clients that you’ve empathised with their concerns and pain points, it will also demonstrate that you are forward-thinking, values-led and, most importantly, unapologetically human.

*Author’s note: I used Google Gemini to help me brainstorm ideas for this article, and for an example of a simple protocol for a fictional video company. I asked it to proof-read an earlier draft for spelling, grammar and flow of ideas. Everything else came from my research, and my own brain.