Eloïse Vo
Eloïse Vo is an artist-designer and PhD candidate in the program between HEAD Hes-so Geneva and the EPFL. Though trained as a graphic designer, her work has branched into performance, media installations, publishing and writing, alongside design commissions. By looping back to John C. Lilly’s experiments, her PhD research explores the ‘Echoes of the Dolphin House’ to recall past histories of synthetic intelligence and contemporary environmental conditions of the wetware from a xenofeminist perspective.
“There are hopeful lessons here, to be found in the latent memory of our attempts, failed or not, to communicate with this Universe we found ourselves within.”[1]
The same way alien quest and interspecies communication were in the 1960s associated with the conquest of two foreign territories – the cosmos and the ocean – the contemporary emergence of Artificial Intelligence requires us to reposition ourselves in the dimensionality of planetary scale computation. By exploring the hype and doom narratives which bring these three episodes together, I tend to look at these past attempts to recall forgotten alternate futures.
This spiraling narration starts in 1973, when humanity was on the verge of communicating with alien intelligence. The spatial vessel Pioneer 11 was about to be launched, as well as the first manufactured object sent beyond our solar system. Preparing for a space encounter, Carl Sagan and Frank Drake, both scientists and pioneers of the Search for Extra Terrestrial Intelligence (SETI), supervised the design of the Pioneer Plate. Made with anodized aluminum, the plate has been attached to Pioneer 11 in order to address a message to this Other form of intelligence. Drawn by the artist Linda Salzmann Sagan, the illustration was designed to figuratively indicate the coordinates of planet Earth alongside the display of many mathematical equations and a representation of humanity: a man and a woman standing. Carl Sagan and Frank Drake were preparing this alien encounter for many years as pioneers of SETI, but also as members of the Order of the Dolphin, a not-so-secret working group named in honor of the American neuroscientist John C. Lilly, who have set their horizon by claiming since 1961 :
‘Within the next decade or two, the human species will establish communication with another species: nonhuman, alien, possibly extraterrestrial, more probably marine, but definitely highly intelligent, perhaps even intellectual.’[2]
John C. Lilly was referring to an already-encountered alien intelligence: the bottlenose dolphin. To allow this encounter, John C. Lilly founded the Communication Research Institute (CRI) in St-Thomas, in the U.S. Virgin Island. Known as the Dolphin House, what was initially a touristic villa on the seashore has been turned into a laboratory allowing the first experiment of inter-species cohabitation. Margaret Howe Lovatt, a 21 year-old woman, lived in immersion with Peter the Dolphin in order to teach him the English language. For seven weeks, Margaret lived with Peter in this semi-immersed space that she designed regarding both their needs from her suspended bed allowing them to sleep protected from salty sea water, to the lift Peter was using to move between their cohabitation space and the downstairs sea pool.
The doomism we inherit
John C. Lilly was very optimistic on Peter’s ability to develop human language, which would have been the first step for the dolphin to gain knowledge and evolve into a human-like species. In parallel with the possibilities of extraterrestrial encounters, the dolphin, as well as the alien, were the object of fantasized other-than-human models of civilization. Despite some stories of companionship between cetaceans and humans, the historian Thomas Moynihan reminds us of the common associations of the dolphin as being “warlike and voracious”, especially in the work of the French zoologist Frédéric Cuvier, who deemed the wild dolphin a “carnivorous brute” and a “stupid glutton”[3]. While John C. Lilly was advocating for a breach in dolphin intelligence by mastering the human language (aka, English), the dolphin started to be perceived as a potential threat, challenging the hegemony of Humans at the top of the pyramid. In the context of extraterrestrial search, this potential evolution of dolphin intelligence catalyzed the fear of being overshadowed by another species, bringing into further focus the severity of that worst silence of all: humanity’s potential extinction.
Recalling these previous stories of the dolphin as an existential risk sheds a light of déjà-vu on recent debates on Artificial Intelligence. Once again, “we” are facing an existential risk with the emergence of the Singularity or the Artificial General Intelligence. This emergence of “AI doomerism,” as the philosopher Emile P. Torres coins it, is led by Eliezer Yudkowsky, who in a recent TIME op-ed endorsed the use of military strikes against data centers to delay the creation of AGI, calling to “shut it down”.[4] AI experts, public and economic figures who regularly express their concern about AI, have published multiple calls arguing that mitigating “the risk of extinction from AI should be a global priority alongside other societal-scale risks such as pandemics and nuclear war”[5]. Reuniting under the acronym TESCREAL, the various interests of the signatories, Emile P. Torres and Timnit Gebru, rendered visible the ideological continuum between Transhumanism, Extropianism, Singularitarianism, Cosmism, Rationalism, Effective Altruism and Long Termism. Without erasing the specificities of each -ism, the overlapping and interconnected movements that the ‘TESREAL’ acronym encompass all emerged from the failed attempt of SETI to bridge space-communication. Our hopes for an extra-terrestrial answer have faded away while the Pioneer vessel continued its race outside of our solar system, revealing the uniqueness and precariousness of human life on Earth.
We could be surprised to see these calls coming from actors working in the field of industrial AI – or be bored by witnessing this mediatic and marketing agitation around it. But the harm is not only what it directs our attention to but what it directs our attention from. According to Kate Crawford, planetary computation could lead us to an “energy crisis” [6]. This noise only reveals the silence of these signatories on the challenge of climate change and current 6th mass extinction, bounding the future of computation with its energetic and environmental dimension.
After the stochastic parrot
Despite his efforts to accommodate this inter-species experiment, John C. Lilly’s attempt of ‘bilateral conversation’ didn’t occur. Many reasons could be listed here to advocate the already-there promise for failure of the “worst experiment ever”[7]: from his self-experiments with LSD to animal abuse. But the main critics addressed his misconception of language learning based on mimicry and repetition, ignoring the biological limitations of the dolphin to articulate most of the necessary phonemes. Almost 60 years later, the same critics are used to temper the excess of fear and fascination regarding Artificial Intelligence, whose use has gained in visibility with the public release of Large Language Models (LLMs) such as Chat GPT 4. AI is often dismissed as a new iteration of the stochastic parrot, mimicking intelligence and subjectivity by using language as subterfuge for their iterative process. Yes, generative AI such as Stable Diffusion, Sora or MidJourney work through imitation and recursion. They are designed so that, when given a set of data (text, image or audio), they will look for optimal statistical solutions of pixels, letters or sounds which present the same properties as the dataset.
AI critics have used parrots as a metaphor for nonhumans who can’t genuinely think but can only spit things back – such as Peter the Dolphin, reviving a presupposed exceptionality of human intelligence – a concept that is everyday more and more difficult to grasp. Beyond this reflectionism, there could also be a deep biomorphism that may correspond to the way humans and animals cognite (even if it’s not how they believe they think). According to Benjamin Bratton and Blaise Agüera, “researchers in animal intelligence have long argued that instead of trying to convince ourselves that a creature is or is not “intelligent” according to scholastic definitions, “it is preferable to update our terms to better coincide with the environmental pressure”[8]. The iterative prediction principle does correspond to an iterative prediction model of biological neurons which is corresponding to the most common way our embedded human intelligence is functioning. Imitation may be found in moments of mimetic synthesis of human and machinic communication, as it opens the door to analysis, to a reflective gaze in the mirror, and also at times, to the surprise of unexpected combinations, some plausible, some incongruous. Thus this deep anthropomorphism shouldn’t be used to discredit the potentialities of AI but rather to re-orient our conceptions from the artificial to the synthetic underlying reality from which they emerge[9].
The question of “intelligence” sounds always-already obsolete – obfuscating the functioning of Artificial Intelligence as a statistical induction based on pattern recognition. Contemporary computation and AI are now functioning under the threshold of human consciousness. Thus, the reduced scale of the Dolphin House should help us to understand the process of externalization of thought. As a cognitive assemblage, the Dolphin House contributed to control and organize the management of information and communication, carried out as much by the human operators as by the dolphins and machines. Saturated with telecommunications technologies such as hydro mics and magnetic tape, the laboratory acted as a media apparatus allowing the close-observation and monitoring of Peter the Dolphin. His vocalizations were recorded on magnetic tape and slowed down, as John C. Lilly considered the difference of speech between human and dolphin needed to be negotiated and matched. Reassigned to the role of mother and teacher of the Dolphin, Margaret Howe performed as a mediating interface in this cybernetic-architecture while her femininity was technologized to benefit the experiment. Both differentiated as less-than-human or more-than-human, the Dolphin House enmeshed Margaret Howe and Peter the Dolphin as embedded components of this “cognitive infrastructure”, as defined by Benjamin Bratton, “creating the material conditions for the synthetic augmentation of the forms of intelligence that emerge in and as complex ecological niche”[10].
The photographic archives of Margaret Howe inhabiting the uncanny valley of inter-species relation became the visible representation of the erroneous anthropomorphism, while preventing us from adopting another perspective. On the contrary, mapping the Dolphin House as a “cognitive assemblage” allows us to reframe the dimensional and relational conditions of “species-in-cyber-symbiosis”[11], as defined by Katherine Hayles. Without erasing the specificities of embedded situations, the Dolphin House prefigures the merging of the biosphere and the technosphere. The mediatic drama around the failure of John C. Lilly’s experiments shouldn’t cast a shadow on the importance of the experiment as a cross-section for many scientists in the 1960s, such as Gregory Bateson, Lois Bateson, or Carl Sagan. By coming back to its reduced scale, the Dolphin House allows us to reformulate our initial question: by moving away from a human-centric conceptualization of ‘intelligence’, this necessary detour generates alternate potentialities of contemporary computation.
Navigating the cosmos, the ocean and the latent space: a xenophilic perspective
While the context of the Cold War accelerated the conquest of two foreign spaces (the cosmos and the oceans), the research of John C. Lilly revealed a noisy and inhabited ocean. The necessity of militarizing and monitoring underwater space and communications reshaped the misconception of the ocean as a hostile and empty space to reveal the vibrant sea life we keep on acknowledging today. The dolphin may not have answered the way we were expecting, but their undeniable presence forced us to engage our “response-ability”, as framed by Donna Haraway.
Thanks to generative AI, our dialogue with cetaceans have continued to improve. Since 2018, the Earth Specie Project has used AI to prevent collisions between whales and other cetaceans with commercial boats. Trained to recognize the language pattern used by whales to signify a danger or a threat, AI redirects these warning signals to the whales when they are approaching a boat. If this project prevents many cetaceans from dying by collision with boats, it integrates the whales as actants of our planetary logistical systems. Thus, planetary computation points to an increased dimensionality of coexistence, by exponentially producing multiple vectors of relations of which humans are not always a part. As Katherine Hayles develops “if machines communicate more with each other than with us, the intervals and pervasiveness of machine autonomy increase – areas where machines make decisions that affect not only other machines but also humans and non-humans enmeshed in this global cognitive assemblage.”[12] This accidental nature of planetary computation requires us to reshape our expectations for a unique Artificial General Intelligence. We need to move away from the spectrum of Singularity prophesied by TESCREAList, in order to acknowledge the overlapping and heterogeneity of the technosphere which have outgrow our individual body scale.
By rendering worlds beyond our perceptual reach, generative AI is producing an estrangement effect, decentralizing the human as alien among other aliens within an alien landscape. We need to move away from the human-centric conceptualization of artificial intelligence which only aligns its computational capacities and power with the modern project of exploitation and alienation of the differentiated others. The Pioneer Plate and its representation of humanity as a heteronormative couple, standing alone in the middle of a mathematical equation needs to be rendered obsolete to engage our “response-ability” regarding this new dimensionality. Planetary scale isn’t just an external condition but also provides a conceptual opportunity to reposition ourselves and our agentivity in the cosmos we inhabit.
Looping back
Knitting together the search for extraterrestrial life, interspecies communication and artificial intelligence only reveal the generative loop where the future is exhausted by new versions of old sameness. AI is an instrumental technology whose effects and materialities need to be critically analyzed, but we also need to dis-align it from its human-centric conceptualization if we want to escape from the binarity of the hype and dooms loops. While our collective memory is encoded in datasets, we need to dive in the latent space to recollect past and alternate histories of coexistence in order to activate their latent potentialities.
If the future has already been canceled, it already has been saved. The same year John C. Lilly was claiming for an alien encounter, Leo Szilard, one of the engineers of the atomic age, published his sci-fi novel The Voice of the Dolphins (1961). The story revolves around a young physicist named Hal Bregg, who discovers a way to communicate with dolphins using a sophisticated computer system. The dolphins, with their recently spoken-language skills, warn humanity from its destructive tendencies, particularly regarding nuclear weapons and environmental degradation. Deriving from the Greek delphus meaning the matrix/the womb, the dolphins are our “adelphe”, our “extra-terrestrial doppelgänger” – whom we recall as a technological mediator every time we feel at risk – and from which our future could happen again.
Références
BRATTON Benjamin, “After Alignment: Orienting Synthetic Intelligence Beyond Human Reflection”, conference at Platform Theatre at Central Saint Martins, 28/06/2023, [online]
AGÜERA Y ARCAS BLAISE, BRATTON Benjamin, “The model is the message”, Noemalag.com, 12/07/2022, [online]
CHATONSKY Grégory, “Intelligence Artificielle et extinction”, AOC.fr, 06/2023, [online]
HAYLES Katherine, “Three Species Challenges. Toward a General Ecology of Cognitive Assemblages”, in LINDBERG Susanna and ROINE Hanna-Riikka eds., The Ethos of Digital Environments, Routledge, 2021
MOYNIHAN Thomas, “Thanks for all the fish”, aeon.co, 03/2021, [online]
BRYLD Mette, LYKKE Nina, Cosmodolphins: feminist cultural studies of technology, animals, and the sacred, Zed Books, 2000
REED Patricia, “Orientation in a Big World: On the Necessity of Horizonless Perspectives”, e-Flux, n°101, 06/2019, [online]
REED Patricia, “Xenophily and Computational Denaturalization” in Artificial Labor, e-Flux, 09/2017, [online]
SAGAN Dorion and Lynn Margulis, “Gaia and the Evolution of Machines”, Whole Earth Review, n°55, Sumer 1987
TORRES P. Emile, ‘The “TESCREAL Conspiracy Theory” Conspiracy Theory”, xriskology.medium.com, 15/10/2023 [online]
Center for AI Safety (CAIS), “”Statement on AI Risk””, safe.ai, 05/2023 [en ligne]
[1] MOYNIHAN Thomas, “Thanks for all the fish, aeon.co, 03/2021, [en ligne]
[2] LILLY John, Man and Dolphin; Adventures of a New Scientific Frontier, Garden City, NJ: Doubleday, 1961
[3] MOYNIHAN Thomas, “Thanks for all the fish”, aeon.co, 03/2021, [en ligne]
[4] TORRES P. Emile, ‘The “TESCREAL Conspiracy Theory” Conspiracy Theory”, xriskology.medium.com,
15/10/2023 [online]
[5] Center for AI Safety (CAIS), “”Statement on AI Risk””, safe.ai, 05/2023 [en ligne]
[6] CRAWFORD KATE, “Generative AI’s environmental costs are soaring — and mostly secret”, nature.com, 20/02/2024 [en ligne]
[7] HOWE Margaret, dans The Girl Who Talked to Dolphins, realized by Christpher Riley, BBC, 2014
[8] AGÜERA Y ARCAS BLAISE, BRATTON Benjamin, “The model is the message”, Noemalag.com, 12/07/2022, [en ligne]
[9] BRATTON Benjamin, “After Alignment: Orienting Synthetic Intelligence Beyond Human Reflection”, conference at Platform Theatre at Central Saint Martins, 28/06/2023, [online]
[10] Ibid.
[11] HAYLES Katherine, “Three Species Challenges. Toward a General Ecology of Cognitive Assemblages”, in LINDBERG Susanna and ROINE Hanna-Riikka eds., The Ethos of Digital Environments, Routledge, 2021
[12] Ibid.
JOURNAL OF THE FIRST OUTERNATIONAL 