Mycelium

LLMs like ChatGPT and Claude are shaping how millions of people think, and will change how entire industries operate - including industries that affect animals at enormous scale, like factory farming. Right now, these models inherit the same human biases toward animals, the moral blind spots inherent in training data.

Why does this matter? If today's models treat animal suffering as indifferent, this bias could shape how an entire generation thinks about animals, and how the systems that perpetuate animal suffering, for years to come.

Mycelium is working towards a solution. We build the tools that measure how AI models reason about animal welfare, then share what we find directly with the companies building these systems, so they can improve their models. These tools (called benchmarks) open the doors towards LLMs with greater moral consideration of animals.

We've already begun building these benchmarks, and we've worked with top researchers in the field, producing published papers on AIxAnimals. We collaborate with organizations like Sentient Futures, CaML, and Electric Sheep.

Your donation helps us: build the tools that test how AI models reason about animals, share what we learn directly with companies like Anthropic, Google, and OpenAI, and help influence the data that teaches AI systems to take animal suffering seriously.