Machine learning is driving rapid progress in “decoding” animal communication – with the potential to radically transform our relationships towards animals. This session explores the ethical and legal implications of such breakthroughs. Should we talk to animals at all, and if so, what ethical considerations should guide our interactions? Which animals are we giving voice to? How might our world change if breakthroughs allowed us to understand animals on their own terms? Are there limits to how much we can understand one another? And perhaps most importantly: Are we ready to listen?
🧩 Central questions
The ethics of contact: Just because we can talk to animals, should we? If so, what ethical considerations might govern how we communicate with them?
Which animals are we giving voice to? How does our relationship with an animal – as a companion, a commodity, or a wild being – shape the ethics of communicating with them? Could advances in interspecies communication benefit some types of animals but not others?
The ethics of listening: How might our moral and legal obligations towards animals change if we could truly understand them? Does an animal's capacity for complex communication change what we owe them, or does it simply make it harder to ignore the duties we already have?
Societal transformation: How would our legal systems and moral codes need to evolve in a world where we could truly understand what animals think and feel? What are the most plausible positive and negative outcomes that could result from complex interspecies communication?
🧭 Learning objectives
Understand: Explain how machine learning is accelerating research into animal communication.
Assess: Anticipate how human society might adapt to advances in interspecies communication, and identify positive and negative outcomes for animals, including backfire risks.
Reason: Explore ethical frameworks and key considerations for interspecies communication (e.g. major implications for human society). Compare and contrast implications for different animal groups (e.g. companion vs. farmed vs. wild animals).
Next steps: Identify priorities across research, policy, and other domains to ensure the ethical use of interspecies communication.
💡
Use the table of contents on the right to quickly navigate this page.
Resources
‣
Required readings
Please review all of these resources prior to your session.
Max Taylor (2024) | 4 min read (stop at § Resources)
This short blog post provides an overview of the use of machine learning in deciphering whale communication. It explains the process, from initial data collection to supervised vs. unsupervised machine learning approaches.
Max Taylor (2024) | 5 min read (stop at § Resources)
This continuation of the above blog post shifts from the science of interspecies communication to the complex ethical questions raised by this technological progress. It explores four potential futures: one where communication fails, and three where it succeeds – raising the crucial question of whether success would ultimately help animals, harm them, or change nothing.
‣
Ethical challenges posed by interspecies communication
Breakthroughs in interspecies communication could spark a new Copernican revolution, fundamentally shifting our perspective from the center of the universe and challenging us to finally learn to listen to other sentient beings with whom we share our planet.
⚠️
When exploring animal communication, it’s easy to fall into two common traps:
Anthropomorphism: Assuming that other animals think and feel like us (e.g. can reason, grasp abstract concepts, communicate, or experience emotions like we do).
Anthropodenial: Assuming that other animals are fundamentally different (e.g. lack key cognitive capacities/intelligence, etc.).
In short: anthropomorphism is projection of human traits onto animals (like when we say a dog feels “guilty”), while anthropodenial is ignorance of traits we share with animals (denying that fish can feel pain).
As you engage with the readings and activities, try to remain mindful of both biases. Strive for humility and a balanced perspective: acknowledging both our similarities and our differences.
Airen (2024) | 16 min read (24 min audio available on Substack mobile)
The present thinkpiece frames both anthropomorphism and anthropodenial as manifestations of anthropocentrism. Our tendency to think of humans as exceptional can lead us to invent false similarities to animals when it benefits us while simultaneously ignoring genuine (yet inconvenient) similarities. Moving past anthropocentrism requires understanding animals on their own terms.
In this article, César Rodríguez-Garavito, Founding Director of the More Than Human Rights (MOTH) initiative, examines the legal and ethical frontier of machine learning-assisted interspecies communication. The author introduces MOTH’s PEPP ethical framework (Prepare, Engage, Prevent, Protect), prioritizing precaution and animal interests over human goals.
Earth Species Project co-founder Aza Raskin shares how AI is already helping us to better interpret other species, offering significant potential for animal welfare.
However, he cautions against risks like exploitation and ecological sabotage. Now is the time to develop ethical frameworks so this technology amplifies nonhuman voices rather than silences them. Indeed, the key challenge isn’t simply cracking the code to talk to animals, but cultivating our capacity to truly listen.
‣
Further reading (optional)
📚
Looking for more? Check out the archive of media and texts related to interspecies communication curated by Earth Species Project and Interspecies.io.
An interdisciplinary initiative applying machine learning to decipher sperm whale “codas”. Marine biologists and linguists collaborate to create a translation interface, with potential implications for cetacean rights & conservation. Access their open-source tools on Github.
A $10 million award offered to research that significantly advances two-way interspecies communication. Finalists of the 2024 competition include work on nightingales, marmosets, dolphins, & cuttlefish.
This interdisciplinary workshop series convenes researchers working in the field of interspecies communication. Experts come from a variety of backgrounds, including ethology, machine learning, linguistics, psychology, and more. Projects span whales, elephants, birds, bats, and even plants– with implications for how we understand ourselves as humans.
This workshop, organized by Interspecies Internet and the Santa Fe Institute, brought together experts across AI, animal behavior, linguistics, and beyond to develop criteria for successfully decoding animal communication. The goal was to establish a rigorous framework for demonstrating translation, hoping that breakthroughs in this area, potentially accelerated by AI, would fundamentally change how animals are perceived and treated.
Communication is more than just through sound and text. In this talk, John Honchariw, founder and CEO of Companion, shows how developments in computer vision enable systems to better understand how animals communicate through movement and posture.
Sara Keen, Senior Research Scientist at Earth Species Project does a live demo of their latest bioacoustic decoding model, NatureLM-audio, illustrating how these tools can analyze vast amounts of animal vocalization data across species.
This blog post summarises the results of a first-of-its-kind global survey (full report here), which found strong public interest in understanding animal thoughts/feelings and widespread belief in animal emotion, language, and culture. While optimistic about AI's potential, respondents expressed concerns about misuse and favored strong regulation, with significant openness to including animal perspectives in governance, especially regarding environmental and agricultural policy.
AI-driven efforts to decode sperm whale communication could set off sweeping paradigm shifts across animal rights law. By translating whale vocalizations–potentially revealing distress over noise pollution or cultural bonds– researchers aim to strengthen legal protections, challenge human exceptionalism, & advance arguments for cetacean personhood. Interspecies communication may well pave the way for the participation of non-human animals in governance.
César Rodríguez-Garavito, David F. Gruber, Ashley Otilia Nemeth, &
Gašper Beguš (2025)
Using cetaceans as a case study, this in-depth legal perspective explores how understanding the content of animal communication (using cetaceans as a case study) could directly reshape existing laws and provide crucial evidence for the fight for nonhuman legal personhood.
This talk presents research using interpretable AI that reveals structured, vowel-like acoustic units (“codas”) in sperm whale communication with linguistic features paralleling human phonology – discoveries with potentially wide-ranging ethical and legal implications.
Christopher Berry & Sankalpa Ghose (2025) | 12 min excerpt (18:41-26:28 only)
This talk brings together Christopher Berry, Executive Director of the Nonhuman Rights Project, and Sankalpa Ghose, founder of Alethic Research, to explore the potential for AI to contribute to the legal representation of nonhuman animals, bridging the gap between their interests and their current lack of legal standing.
This book is a deep dive into how digital bioacoustics and AI are revealing the hidden soundscapes of the natural world, from whale songs to plant vibrations. This work showcases how technology can deepen our connection to other species while also highlighting the profound impacts of noise pollution.
This Oscar-winning documentary records repeated encounters between a filmmaker and a wild octopus as they form an extraordinary bond, illustrating the potential for interspecies connection without complex two-way communication.
Virginie Simoneau-Gilbert and Leonie Bossert (2025)
From including animal interests in governance to AI confabulation and black box systems, this blog post traverses the opportunities and risks of interspecies communication. The authors conclude with three principles for ensuring that technologies developed to understand animals are deployed to their benefit.
This discussion explores the ethical, legal, and relational implications of using AI to decode whale communication. Unique insights include César Rodríguez-Garavito detailing the PEPP framework (Prepare, Engage, Prevent, Protect) for guiding research ethically, and Bridle challenging the centrality of language, arguing that true kinship requires care and practice, not just translation.
This article provides a critical perspective on AI-driven efforts to decode sperm whale communication. It emphasizes the limitations of AI pattern-matching without genuine understanding, raising concerns about potential misuse (e.g. cultural contamination, termed by Aza Raskin as “whale QAnon”). The piece also introduces alternative views emphasizing empathy and shared experience (e.g., through music) over direct translation, arguing we already know what whales need: for humans to stop harming their environment.
‣
More on anthropomorphism vs. anthropodenial
“If a lion could speak, we could not understand him.”
“It will not help to try to imagine that one has webbing on one's arms, which enables one to fly around at dusk and dawn catching insects in one's mouth; that one has very poor vision, and perceives the surrounding world by a system of reflected high-frequency sound signals; and that one spends the day hanging upside down by one's feet in an attic. Insofar as I can imagine this (which is not very far),
it tells me only what it would be like for me to behave as a bat behaves. But that is not the question. I want to know what it is like for a bat to be a bat.”
Kristin Andrews & Susana Monsó (2021) | See especially §2.3 Objectivity
This subsection of the Stanford Encyclopedia of Philosophy’s entry on Animal Cognition examines issues relating to the interpretation of research into animal minds. It introduces Morgan’s Canon (we should prefer simpler explanations of animal behaviour), along with other related concepts, such as anthropomorphism, anthropodenial/anthropectomy, and anthropofabulation (the double mistake of overestimating human capabilities, then denying that animals have similar abilities based on this inflated standard).
Primatologist Frans de Waal critiques fears of anthropomorphism, coining anthropodenial* for wrongly rejecting human-animal similarities. He argues that anthropodenial tends to cause more harm than anthropomorphism. De Waal promotes animal-centric**anthropomorphism – applying human concepts informed by the animal’s perspective – as a valid scientific tool for generating testable hypotheses.
This article warns against rehabilitating anthropomorphism as a scientific approach to animal behavior – even in its “critical” or “animal-centric” forms.
This article frames anthropomorphism as an intuitive cognitive bias rooted in unconscious mental shortcuts, similar to social biases. If this is right, then we should manage this bias in the same way we manage other cognitive biases. Namely, Dacey argues that we should develop new strategies based on well-established social psychological protocols, such as using checklists instead of simple rules like Morgan’s Canon.
Please spend 20-30 minutes completing the following three exercises.
You can write your responses in bullet point format if that’s easier.
Submit your responses in the weekly Slack thread created by your facilitator in your channel at least 24 hours before your regularly scheduled meeting.
Leave at least one comment on somebody else’s response.
Which animals do we give voice to?
[150 words]Advances in interspecies communication won’t happen for all species at once. A breakthrough with one group could have vastly different consequences than a breakthrough with another. Our existing relationships with different animals – as companions, commodities, or wild cohabitants – will shape how society reacts and transforms.
Compare and contrast how human society might transform (or not) if we developed a rudimentary means of deciphering key welfare-relevant signals (e.g. vocalizations, movements, or behaviours indicating pain, distress, fear, environmental preferences) with each of the following:
Sperm whales (representing wild animals)
Chickens (representing farmed animals)
Cats and dogs (representing companion animals)
The transformative potential of interspecies communication: a reality check
[150 words]The idea of “talking to animals” is exciting, but will technology alone be enough to create meaningful change? Let's consider two sceptical arguments:
We already know they suffer: We can already infer a great deal about animal well-being. We don't necessarily need complex two-way communication to know that a factory-farmed pig is suffering.
We often don't listen to humans: The capacity for language does not guarantee compassion or action. Humans clearly communicate their suffering, yet exploitation and injustice persist around the world.
Do you agree or disagree with these worries and why?
