As an agency, we recognize the transformative force AI will present to many industries, including our own. This policy addresses how Coldwater Communications will approach AI usage at the agency and on behalf of our clients.
- Human oversight: Any usage of AI, machine learning, and data usage requires human oversight. AI tools may be used to help enhance and augment human intelligence. All outputs generated by AI must be verified, edited, and fact-checked by humans.
- Transparency: Usage of AI, machine learning, and data usage on behalf of clients or outputs generated through AI tools should be disclosed to clients. Coldwater may disclose the use of AI tools for workflow and research-related tasks should the client ask. Research sources and citations should always be documented should a client ask for more details.
- Data Security and Privacy: Coldwater Communications will do its utmost to ensure clients’ data security and privacy are protected to the best of its knowledge and ability. Sensitive and proprietary data will not be shared on AI tools and platforms that the client has not screened and consented to. Closed enterprise-level platforms are preferred.
- Quality of work: Coldwater Communications remains dedicated to providing our clients the highest quality of work and care. Strict quality control will be adhered to when reviewing outputs generated by AI tools for plagiarism, inaccuracies, errors, and any copyright infringements.
- Ethics and Biases: Coldwater Communications recognizes biases exist in the algorithms that operate AI. We will do our best to review AI tools, understand what biases may exist, and counter it through more specific and diverse prompts. We will ensure any usage of AI-generated outputs does not conflict with our values as an Agency, and that we are not creating harmful outputs that perpetuate discrimination, are biased, offensive, or create mis and dis-information.
- Third-Party AI Usage: We will require third-parties (such as influencers) on client work to disclose use of AI in generative content.
- Copyright: Take copyright into consideration when using AI generated output. Do not use AI generated imagery, music, or art in any final client output due to potential copyright infringements.
- Training and Education: We will remain dedicated to ongoing training and education around AI and data usage for our employees and encourage the flexibility to experiment and test tools and new technologies as they emerge in a safe way that complies with our AI policy and principles.