AI Policy for Video Production

MiniMatters AI Policy

As a video production company, we recognize that the growth of artificial intelligence (AI) will have a profound impact on our work and a business that we have built by producing high-quality videos and though developing a relationship of mutual trust with our clients over many years. Furthermore, our clients largely include nonprofit organizations and public or quasi-public agencies. As the stewards of philanthropic and public funds, transparency and trust are important to our clients, as well. For this reason, we articulate in this policy how MiniMatters will approach the use of artificial intelligence in our work.

AI Policy for Video Production that Focuses on Generative AI

We focus this policy on Generative Artificial Intelligence (called “GenAI”) used by applications like ChatGPT (by OpenAI), Gemini (by Google), Claude (by Anthropic), Microsoft Azure, and Eleven Labs. These applications scan or are trained on bodies of existing work, material or data, and can then produce words, phrases, code, images, sounds, voices, music, videos, etc. using statistically-based models. When we refer to AI in this policy, we are referring to GenAI.

AI Policy for Video Production that Is Based on Our Principles

We want our AI policy to reflect our values, which we articulate in the principles below. Furthermore, we recognize that AI is changing rapidly and that we will not be able to anticipate the many ways it will affect our business. However, these principles provide us with a framework for decision-making.

Principles about Video Production and AI

• We will think for ourselves first and in earnest. We are a creative company that produces videos and stories for things that matter to the world and make it a better place. As such, we believe that humans should be at the center of our creative process. To us, this means that our creative direction and initial processes should be done by humans, and not generative AI.
• We will prioritize having creative work done by people. We support creative and other workers who bring skills, intelligence, values, and purpose to their work, and whose livelihoods and families’ well-being depends on the income that comes from their work.
• We will never use AI with the intent to deceive. We will not use AI for deceptive practices, such as making someone appear or sound as they’re not, or making someone say something they did not. This has been and will continue to be our policy with non-AI-based editing as well.
• We will respect the energy needed to power AI and environmental impacts. We recognize that AI is using vast and increasing amounts of our planet’s energy. Because that energy use has major costs and environmental impacts, we will aim to use AI only when it advances our business purposes consistent with our other values, and not for frivolous purposes.
• We will respect privacy and intellectual property. AI applications work based on models that are trained on existing materials/data/information that may be the intellectual property of creators and may have been used without permission and/or compensation. Like with environmental concerns, we will, therefore, factor this into our choices about when and how to use AI. Furthermore, when people use AI, the default setting on many AI applications is to allow data input from the user (such as prompts, documents, images, etc.) to be used to further “train” the AI model. However, if we use AI, inputs may be the intellectual property and ideas of our company and our clients, some of which we or our clients may not want shared publicly. To respect that and stop AI applications from using such proprietary information, we will turn off such “use my information to train the model” features.
• We are aware that AI can be uneven in the sources it draws upon. Therefore, the results it generates may under- or over-represent certain voices, facts, or points of view. We will be cognizant of this in determining when to use AI and when evaluating AI content we generate.
• We will endeavor to verify information we gather from AI. We will attempt to verify the information we obtain from AI sources. This process may include steps such as checking sources to make sure they are credible, comparing information to our own body of knowledge, or having more than one person review the AI-generated information.
• We will use AI with the intent to improve to quality of the videos and other products and services we produce. We wish to continually improve the products and services we provide, including videos and related products, such as transcripts, captions, and still images. Where use of AI does not clash with other principles or goals, we will take advantage of these capabilities where we believe they can bring improvements.
• We will use AI to streamline the production process and achieve efficiencies. AI offers a wide range of business tools, whether related to video or not, that will enable us to do our work more efficiently. Where AI use does not clash with other principles or goals, we will take advantage of these offerings.

Specific AI Uses and Practices for Video Production

Based on our principles, we articulate below some of the main ways MiniMatters currently uses or anticipates using AI tools:

• Voiceovers. Occasionally, and only based on prior discussion with the project’s client, MiniMatters will generate voiceovers using AI. We typically do this in situations when we reasonably expect late-in-the-process script changes that would otherwise result in needing to re-record a script on tight turnaround, or where a consistent voice is needed across projects and where the same artist might not be available. Otherwise, we prefer to hire professional voiceover artists. We use AI voiceovers only in tandem with original content.
• Audio Editing. The most common professional video editing software programs include features to improve quality. Many of these editing features do not use AI to alter the voice. However, AI now provides additional capabilities, for instance, isolating and removing certain sounds (such as background noise or hums). We believe that utilizing these capabilities improves the final video viewer experience without changing the content, substance, intent, or identity of the speaker. Therefore, we employ AI audio editing capabilities. We do not generate speech that a person we interviewed did not say.
• Visual Enhancements. Video editing software used by MiniMatters includes AI features that allow the video editor to improve visual appearances. This includes capabilities such as color correction, brightness/contrast adjustments, or skin smoothing. While many of these software capabilities have been around for a long time and may not necessarily be based on generative AI, we anticipate that the capabilities will increasingly rely on AI and will improve technically. Like with audio editing, we believe that utilizing these capabilities improves the final video viewer experience without changing the content, substance, intent, or identity of the speaker.
• Backgrounds and Placing Speaker on New Background. AI is improving the capabilities to separate a speaker from the background in front of which they were originally recorded. This process can go by various names, including greenscreening, masking, rotoscoping, cutting out, or background removal. AI may also be used to extend a background. We use these techniques in various types of video projects when it serves a strategic and intentional communications purpose, and with the understanding of the client and where there is no intent to deceive.
• Captions. Generative AI is streamlining the production of captions and is, therefore, making video content more accessible. We feel this is a positive use of AI and will use AI-captioning in our videos. However, we also provide a human check on the captions to ensure their quality.
• Graphics. When we are creating an animated video, we need to create or license vector graphics, which we then animate with motions and movements. While we generally strive to create these graphics ourselves or license them, this is an area where we feel generative AI can occasionally be helpful in creating specific graphics faster, more efficiently, or more flexibly, while adhering to our general AI-use principles.
• Photos and Video Footage. AI-generated photos and video footage is an area where we believe care needs to be taken in determining whether to use AI, because of the risk of communicating inaccurate information. We will not use AI-generated photos or video footage with an intent to deceive. We would consider using AI-generated photos and video footage in limited situations and with special care to ensure that they do not compromise the credibility of the information we’re presenting. In addition, we will communicate up front with our clients about their own AI policies and any AI-generated content being considered for a video project, so that they can have agency in the decision and so that they can disclose, when necessary, that AI-generated content was used. Overall, though, in keeping with our principle to prioritize having creative work done by people, we will prioritize the use of actual photographs and video footage.
• Writing. Writing encapsulates a very broad range of potential AI uses. For video production, the main way that AI is playing a role is in script writing. However, MiniMatters believes that we, as humans, should write video scripts for our clients when part of a project scope. This includes the initial drafting, as well as the first several rounds of review. After that, in some cases, we do believe that it could be appropriate to check a script using AI for such uses as shortening the script or modifying the reading-level complexity of a script.
• Other business uses. There are numerous other ways we might use AI for our business, including such tools as Zoom/Teams meeting summaries, email organizing/summarizing management tools, and AI tools embedded in our financial management software. For these uses, we will first determine if there’s a compelling business use for using the AI capability that outweighs other factors such as human capacity for comparable thinking, energy use, etc. If we determine that there is, we may use the AI tool, but will also carefully review the result/output ourselves. We will write our own emails, thus helping our clients know that they are getting our own thinking.

This is not an exhaustive list of how AI is affecting our business and video production, but it includes the main ways MiniMatters is using AI to date.

Sharing Our AI Policy for Video Production

We will provide our AI policy to our contractors. This will help ensure that our contractors are aligned with our policy. We will reference our AI policy in our Standard Terms and Conditions in our agreements with independent contractors and will call attention to this change. We will ask contractors to disclose to us if they would like to consider using AI in a way that is known to require disclosure on video-hosting platforms. So long as it does not conflict with our other policies or practices, or those of our customer/client, we encourage our talented pool of professionals to choose when and how to most appropriately use AI tools to do their work better,

Looking Ahead

We will adapt this AI Policy as needed. If you have thoughts to share about it, we welcome your input.

And, for the record, this policy was written entirely by our team at MiniMatters, with no assistance or input from AI.

Published September 17, 2026