When Whales Speak: Using AI To Communicate

The Earth Species Project is building AI models to decode the languages of whales, crows, and elephants. If it succeeds, the implications extend far beyond biology: it could ignite a legal and political battle over non-human personhood. Who gets rights — and who gets to decide?

From Science Fiction to Field Lab

For decades, the idea of “talking to animals” was confined to science fiction. Today it is a research program. The Earth Species Project (ESP), a non-profit founded in 2017, is attempting to build large language models not for English, Chinese, or Spanish, but for the sounds of whales, elephants, and birds. Their mission is stark: to illuminate non-human intelligence through AI.

It is an ambition that could redefine the boundary of personhood itself. If a machine can reveal that whales have structured communication, how long before the question is raised in court: does a whale deserve rights?

From Listening to Speaking

ESP’s flagship system, NatureLM-audio, is designed to ingest massive datasets of animal sounds. The goal is to detect patterns that may constitute vocabulary, syntax, or cultural variation. The team has gone further: it is now generating synthetic calls. In controlled experiments, AI-created whale songs and bird calls are played back to test whether real animals respond.

This is not just observation — it is intervention. Once machines “speak back,” science risks becoming manipulation. The line between decoding intelligence and creating it dissolves. At that point, the ethical stakes change. Are we studying communication, or are we colonising it?

Personhood: From Elephants to Algorithms

ESP is not operating in a vacuum. The Non-Human Rights Project has already fought cases in U.S. courts for chimpanzees and elephants, arguing that their autonomy and cognition justify legal personhood. Courts in Argentina, Colombia, and India have already granted legal rights to rivers, elephants, and monkeys. The precedent is expanding.

If ESP’s research demonstrates structured animal communication, the argument strengthens: intentional communication implies interests, and interests imply rights. The debate that began in zoos and sanctuaries could spill into international law.

But here lies the danger. Granting rights without corresponding duties creates liability vacuums. If a whale has legal standing, can it sue? If an AI agent has “continuity rights,” who pays when it causes harm? The debate risks being hijacked by corporations that want legal personhood for their AI systems to protect profit while dodging accountability.

The Jungle of Intelligences

A scientist once remarked that the threshold of self-awareness seemed to emerge at the Crab. Mosquitoes do not know they are mosquitoes; dogs clearly know they are dogs. Somewhere in between, a Crab may glimpse itself. ESP’s work could sharpen that boundary — or redraw it entirely.

At the same time, humans are already projecting agency onto machines. Millions of people treat Alexa, Siri, or ChatGPT as conversational partners, often with politeness. Politeness is not proof of consciousness, but it is proof of social effect. ESP’s breakthroughs will amplify this tendency. If whales respond, how many will insist they “speak” in the human sense?

This is the jungle of intelligences: biological, artificial, collective. “Borganisms” and prediction markets may emerge alongside whale clans and corporate AI swarms. The risk is not only misrecognition, but reallocation of moral and legal weight across species and systems.

Economic and Political Stakes

The implications are not academic. If whales, elephants, or corvids are recognised as communicative actors, industries face disruption. Industrial fishing, whaling, deforestation, and factory farming could face new legal challenges. A “communication right” for animals might be used to halt projects that interfere with their habitats.

Corporations will not watch passively. The same arguments could be repurposed: if a crow has rights, why not a corporate AI agent? If animals can “own” their communications, why shouldn’t AIs “own” their outputs? This is how liability shifts: from human overseers to legally shielded non-humans, leaving humans with responsibility gaps but corporations with control.

Cautionary Notes

There are scientific, ecological, and legal pitfalls. Animal “languages” are not like human languages; decoding them risks projection. AI-generated calls could disrupt species behaviour. Expanding rights without duties creates incoherence. And in politics, personhood debates are always vulnerable to capture by activists or corporations with their own agendas.

What begins as a quest for understanding can quickly mutate into a tool of control. Decoding is never neutral. Whoever defines the meaning of non-human signals defines their value — and claims the power to speak on their behalf.

Closing Verdict

The Earth Species Project could open a door we cannot close. Once whales “speak,” society will face a legitimacy crisis: do we treat them as partners, petitioners, or merely data sources? The rhetoric of liberation may mask a scramble for power, as corporations and campaigners alike rush to fix definitions in law.

The question is not whether whales, crows, or AIs “deserve” rights. It is who decides the hierarchy of intelligences — and in whose interests. The jungle of minds is coming. The real predators will be those who control the definitions.