Could Chat GPT Talk to Whales?

This is the resonance emanating from the largest-toothed apex predator existing on our planet, a melodic broadcast emerging from the unfathomable depths of the sea – a call that diverges from harmonious melodies characterizing other whales' vocalizations, manifesting more akin to digital data transmissions. This auditory phenomenon has reverberated through the annals of nautical history, echoing through the centuries, ensnaring sailors' imaginations with its eerie wails, erroneously believed to be the lamentations of drowned mariners' spirits. Astonishingly, it wasn't until 1957 that the scientific community realized these sounds originated from cetaceans, and it took until the 1970s to comprehend the complex nature of these undulating clicks and bellowing utterances as a form of communication.

In the contemporary milieu, scientists now harbor the belief that sperm whales could hold the key to breaching an unprecedented barrier: interspecies communication. This cetacean species, they contend, presents a unique prospect for unraveling the enigma of communication between disparate species. If one were to place a bet on any creature harboring language akin to the human variety, the sperm whale would emerge as a frontrunner. Over time, researchers have gleaned insights into the prodigious development of their brains, their intricate social and familial structures, and a cultural framework held together by a sophisticated communication system.

In many respects, the lives of sperm whales bear intriguing parallels to our own, sparking speculation that their language might have evolved for similar purposes. These leviathans have wielded their massive brains for tens of millions of years longer than the span during which our modern-sized brains developed. With their extensive repertoire of distinct sounds and intricate patterns, the language of sperm whales emerges as a realm of potential complexity akin to our human languages.

Human yearning for conversing with other creatures has long been a facet of our collective imagination, entwined within folklore, fables, and fantasy across various cultures. In an unprecedented juncture, we find ourselves equipped with technology that might finally bridge this chasm between species. Recent years have witnessed significant strides in natural language processing, epitomized by endeavors like ChatGPT3, providing a glimpse into the potential of computers engaging in human language. As this technology advances, scientists perceive an opportunity to apply these principles to the realm of interspecies communication, with the sperm whale as a captivating subject of study.

The dawn of the Citation Translation Initiative (SETI) in March 2020 ushered in a pioneering effort in interspecies communication. This ambitious project centers around the collection of data from a community of sperm whales dwelling in the vicinity of Dominica, a Caribbean island. These whales have undergone over a decade and a half of scrutiny under the aegis of the Dominica Sperm Whale Project, setting the stage for an endeavor of unprecedented scale.

Machine translation, a field that has revolutionized human language, serves as a potential linchpin in deciphering the cryptic language of these marine giants. Since 2014, machine translation has hinged on the utilization of encoder-decoder deep learning models. These neural networks work in tandem: the encoder processes a sentence, converting its words into multi-dimensional vectors called word embeddings, while the decoder takes these vectors to generate a translation in another language. However, for this process to function, it demands human supervision and relies on pre-existing pairs of input-output data.

In 2018, a breakthrough emerged from the annals of Facebook Research, demonstrating the viability of unsupervised translation. This innovative approach involves deciphering the statistical properties of words within a language, essentially constructing a three-dimensional "galaxy" where words are positioned based on their correlations and frequencies. This technique permits translation between dissimilar languages, eschewing the need for explicit human guidance, as the structural parallels within the "galaxies" facilitate the mapping of languages.

Tom Muscle, a documentarian and author of "How to Speak Whale," holds a profound interest in SETI's work. He envisions the possibility of unveiling the "shape" of the universe of whale communication, perhaps identifying junctures where their linguistic realm converges with ours, offering a potential window into their language and cognition.

The process of capturing data for SETI's translation efforts is nothing short of intricate and multifaceted. It relies on a consortium of autonomous and semi-autonomous devices strategically positioned to capture a holistic picture of the whales' vocalizations. Tethered buoy arrays, akin to submarine detection systems, encompass audio recording devices deployed at varying depths, tracking the whales' sonic narratives. Recording devices affixed to the whales via suction cups provide unparalleled insight into their vocalizations and behaviors, while aquatic drones autonomously capture audio and video, rectifying potential blind spots.

The colossal volume of data collected necessitates a spectrum of machine learning techniques. Algorithms will sift through the auditory cacophony, distinguishing between different types of sounds – echolocation versus communicative vocalizations, for instance. They will ascertain which whale is speaking, the context of the conversation, and potentially infer meaning from patterns and variations in vocalizations.

Crucially, SETI researchers must grapple with the challenges of interpreting these vocalizations. Sperm whale codas are intricate and multifaceted, with variations in amplitude, frequency, and structure, raising questions about the significance of these nuances. The quest for translation also raises philosophical inquiries, such as whether translation presupposes shared cognitive and emotional states. Moreover, the whales' unique perspectives of time, self, and emotion may give rise to concepts and vocabulary that defy human comprehension.

While the successful translation of sperm whale language remains uncertain, the endeavor holds immense promise. Beyond the potential to decode their communication, the pursuit itself speaks to our innate connection with the natural world. As ocean ecosystems face peril from climate change, anthropogenic disturbances, and ecological shifts, understanding the intricate lives of these cetaceans assumes paramount importance. The act of striving to comprehend and connect with these enigmatic beings reflects a profound step toward nurturing and safeguarding their existence, perpetuating the age-old human tradition of seeking connection with the creatures that share our planet.