Earlier this year, we released Part 1 of our Artificial Intelligence (AI) series on the current AI landscape as well as Harmony Labs’s past and future-facing uses of AI. Within it, we underscore the importance of taking a human-led approach and using our core values to navigate AI decisions. Here, in Part 2, we share precisely what those values mean for how we use AI in our work.

We Use AI

Technology, advanced modeling, and analysis are core to our work. This means that AI is also core to our work. The resources, methods, and tools we use are continuing to evolve, and we are evolving with them.

We Protect Our Partners

We are responsible for the ethical, contractually appropriate use of the data we own and the data we steward on behalf of partners. We don’t publish or share any models we create to analyze partner data. Except where we are explicitly permitted to do so, we don’t submit any partner data to chatbots, embedding engines, or external AI platforms.

We Protect Our Data

We require a rigorous review and approval of an AI platform’s privacy and data protections prior to using the platform even with our wholly owned data. We will never use a platform that steals, sells, or learns from what we provide it. We stay up to date with policies and terms of service as they change to make sure that our data are secure.

We Employ People

We always employ and pay real humans—writers, filmmakers, artists, software engineers, data scientists, and researchers—a living wage for their work. That means that we do not accept the work, analysis, or judgment of AI without human review and refinement in any of our functions from engineering to analysis to content creation.

We use AI as a tool to generate new code for our engineering codebase. 100% of AI-generated code is modified by a human to ensure it is fit for purpose. 100% of the code committed to our codebase by engineers is reviewed and approved by a second human engineer.

We use AI as a tool to improve our efficiency in understanding complex media, to generate summaries, insights, and new ideas. 100% of AI-generated summaries, insights, and ideas are reviewed and verified, rejected or revised by a human strategist or storyteller before they are published to a partner or the public.

We use AI as a tool to support our design and communications capabilities by generating sketches for creative briefs, visual summaries of media, and novel content. Sometimes creative professionals use AI in their process, and when our team does that, they use it because it is the best tool to achieve their creative intentions. 100% of the content we put into the world in our reports and elsewhere is conceived, created, reviewed, and approved by design and communications professionals.

We Empower Humans

We do not use AI to replace human analysis, storytelling, or advocacy. We use AI to give analysts, storytellers, and advocates more freedom and power to engage with their audiences authentically, creatively, and as humans.

We choose AI applications that facilitate the imagining and creation of new possibilities for how we can work and be together as humans. We reject AI applications that might actively harm humans or constrict possibility. When considering creating something with AI, we ask ourselves: “Will people touched by this product be empowered and enriched, whether as users, audiences of its outputs, or as invisible contributors?”

We Won’t Stop Here

The data and tools we have access to—AI and otherwise—will continue to evolve, and so will we. As stewards, not just of our partners’ data but of all the stories we tell each other and consume, we will continue to hold ourselves accountable for supporting narrative work which creates more liberty, empowerment, and connection for storytellers and their audiences.