Blue Zoo’s Manifesto for the Ethical Use of Generative AI: We Use AI For Good
As a B Corp animation studio committed to leveraging technology for the betterment of society and creativity, Blue Zoo establishes the following manifesto to guide our ethical use of Generative Artificial Intelligence. It reflects our dedication to innovation, integrity, and responsibility.
Our stakeholders (executive team, staff and clients) have highlighted the anxieties, concerns, risks and potential benefits around AI usage, therefore, it is critical we have a clear and concise policy.
Introduction
From our research, we have identified five key categories of risk and concern:
- Legalities and ethics of training data
- Quality control of resulting output
- Resulting job displacement caused
- Environmental footprint of the compute
- Resulting lower job satisfaction
We have formulated 10 statements, which help frame how we approach these five issues in our studio:
We Prioritise Human Creativity
We will use AI as a tool to empower human creativity, not replace it. Our core belief is that AI can enhance the creative process, enabling our artists and creators to reach new heights of innovation and visual storytelling by allowing them to focus on the artistry over the technical process. Our craft will remain originated by humans, including but not limited to, concept design, script writing, animation and voice-over.
We will give our artists the freedom to decide whether or not to use AI tools in their work, and will only ever specifically ask them to do so for experimental and evaluation purposes. Anyone who wishes to use AI tools should and will be free to do so, as long as they recognise and abide by the policies outlined in this Manifesto.
We Continuously Learn and Improve
To ensure we are choosing our tools correctly, we must stay informed about what tools are out there and their potential impact, both positive and negative. This will enable us to adequately consider whether they align with our standards, our clients' standards, and our technical competency.
We will continuously explore the ethical considerations of current and emerging AI technologies to ensure our practices and policies, including this Manifesto, remain aligned with our values. We will routinely share our learning and evaluations internally, classifying AI use cases into a “traffic light” system based on the five areas of concern.
We Do Not Delegate Accountability
Generative AI tools are prone to hallucinate, therefore we will not use AI tools without human oversight, to ensure we remain compliant with our responsibilities, both legally and ethically. AI tools should be used as an aid to help maximise creative work, freeing staff from what they consider mundane, but not substituting or delegating human accuracy or accountability.
We recognise that there is an expected job displacement with any optimisation of a technical process, but we believe such optimisations will only be increasing personal productivity by allowing staff to focus on more impactful areas, similar to any other - non-AI-based - technological automation (e.g. better UV unwrapping and retopologising tools).
We Will Not Utilise Unauthorised Copyrighted Training Data In Production
We are responsible for respecting the rights of creators and artists, both legally and ethically. We may use generative AI as a tool when brainstorming or creating mood boards only to help explore and communicate ideas, similar to how we use copyrighted reference artwork, but we will not use AI trained on unauthorised copyrighted artwork to generate content that will appear in final deliverables.
We recognise that the vast majority of generative image generation AI models are trained on unauthorised copyrighted material. Therefore, to avoid financially supporting that practice we will:
- only subscribe to AI vendors that exclusively train on authorised data (e.g. Firefly);
- run free, locally hosted alternatives with commercially permissive licences (e.g. SDXL);
We will also not use data to train our own models without explicit permission from the copyright holder, including voice-over artists.
We Default To Transparency
Transparency and trust are essential for maintaining strong relationships with our employees, clients and audiences. Therefore it is critical for us to be transparent about how AI technologies are implemented within our projects.
This includes being transparent about where AI is used, how it influences decision-making processes, and the origins of the AI technologies we adopt. Any content presented to clients utilising AI-generated content, such as mood board images, will always be clearly marked as such.
We will also maintain an internal register of the generative AI tools we use and how they are applied within our projects.
We Minimise Our Environmental Footprint
Our studio has a science-based approach to net-zero goals, through carbon emission reduction over offsetting. Therefore in developing and implementing AI tools, we prioritise solutions that report their environmental footprint and have the lowest carbon emissions and fresh water usage, and always consider if the footprint outweighs the contribution to our impact.
Additionally, we prioritise running AI models on our own or rented infrastructure - local or cloud - where we have full visibility of the energy usage and choice of energy providers.
We Protect Human Privacy and Data
In using AI that necessitates data, we will ensure stringent adherence to privacy laws and practices. We will protect the information of our employees, clients, and stakeholders, making sure that personal data is never misused or compromised, and the tools we use have been vetted for compliance.
We Champion Inclusivity and Diversity
We pledge to use AI technologies to promote inclusivity and diversity. This means not using models that may perpetuate stereotypes or inequalities by restricting the use of AI for job applicant processing or personnel decision-making.
We Use AI For Good
We will only deploy AI in ways that respect the dignity and rights of all individuals impacted by our work, helping aid our pursuit of using animation as a force for good. This includes a commitment to not using AI in ways that deceive or misinform audiences or misrepresent our creative processes, adhering to our ethical marketing and content policy.
We Engage and Collaborate with our Community
Recognizing the broader impact of AI on society, we will engage with staff within our studio and the wider animation community, including other studios, technology companies, and regulatory bodies, to promote the ethical use of AI in animation. We believe in collaboration over competition in addressing the ethical challenges AI presents.
This manifesto serves as a foundational document to guide Blue Zoo in responsibly harnessing AI's power. We are committed to revisiting and revising this manifesto as technology and societal values evolve, ensuring that our use of AI remains ethical and aligned with our mission to use human creativity as a force for good.
Manifesto Compliance
If any staff member is concerned about this manifesto not being respected, please talk to your line manager, a member of the Exec team, or follow our whistleblowing policy.
Generative AI was used to aid this document's structure, formatting and quality control.