Our AI Policy
How eighteen73 is using Artificial Intelligence
We are excited about the potential that AI tools and technologies can bring; providing opportunities to deliver excellent work for clients in new and and effective ways.
However, as with everything we do, we remain thoughtful and pragmatic about how to adopt these new tools into our work. We are conscious of the wider implications to our team, the industry and our environment.
AI should be used to complement, and never replace human creativity, with security and privacy a prerequisite for any tools we use.
In Summary
- We review anything assisted by AI and never mislead clients about its use.
- Personally identifiable information is dealt with extremely carefully.
- Intellectual property and copyright are never given up to AI tools.
- We empower our staff to make their own judgement about where AI is effective, but they work within sensible constraints.
- Security and ethical considerations are at the front of our minds when picking up any new AI tool.
Our Guiding Principlesย for AI
We commit to the following central principles whenever we use AI
Security
We only use AI tools from reputable companies. If the AI company may use our input as a basis for further learning, we do not share PII or any other sensitive information. If unsure how the company will store or use our input, we make our prompt more abstract and anonymise any data.
Our team of developers are aware that any code they produce with the assistance of AI is still their code in every usual way. They remain accountable for its security, functionality and performance. AI is a powerful productivity tool to speed up development but our team understands and checks every line they are assisted with.
Transparency
Where possible, we will always convey to clients when our final work consists of any meaningful use of AI, particularly in creative works. As productivity software incorporates more AI functionality the line is sometimes blurred so we err on the side of transparency when unsure.
Some examples of what we may consider to be non-meaningful works would be the use of AI in routine spelling and grammar checking, summarising documents, transcribing meetings, or generating placeholder content that won’t be used in production.
We never use an AI pretending to be human for client-facing or colleague-facing communications (e.g. customer service agents, email automation) without explicitly identifying them as such.
Team Empowerment
Our approach with AI is to use it to boost our team’s productivity, making the more laborious elements of their work faster and more fun. Itโs a service for our team to utilise and find a professional advantage, not something to replace them.
We are conscious of the impact emerging technologies may have on colleagues entering the industry and the potential for more junior tasks to be replaced with AI. Our agency remains committed to bringing on new talent by recruiting junior staff members and taking on apprentices.
All team members are briefed on our AI policy, and we continually assess any effect it may be having – positive or negative – on their work through open dialogue.
Ethical Concerns
We acknowledge the environmental impact that is an unfortunate side-effect of using AI. There is undoubtedly a big energy cost behind the complex AI operations taking place, and a lot of natural resources are needed to produce AI models.
We ask our team to remain alert for biases that may exist in AI models. Aside from most modelsโ natural tendency to Americanise their output, more dangerous biases could relate to under or overrepresented social groups, weighted social views, or some other inequality we do not want to reinforce.
We consider all factors when choosing AI tools and hope that more users being open about the problems will force suppliers to strive to do better themselves. We also hope our wider efforts to run an ethical and energy-efficient company help to offset some of the cost.
Intellectual Property
The UK government1 is racing to ensure laws encompass issues raised by the rapidly evolving AI industry. Whatever legal refinements are needed, it will be the case that AI-assisted works will have copyright assigned just as before; i.e. the human creating the content owns the IP.
There is a grey area where content is solely AI-generated, but as a policy, we do not deliver such work.
This policy was last updated in January 2025