[ExI] LLM are connecting all terrestrial intelligence

Thu May 21 17:21:42 UTC 2026

On Mon, May 18, 2026 at 5:18 AM BillK via extropy-chat <
extropy-chat at lists.extropy.org> wrote:

> On Mon, 18 May 2026 at 04:32, Stuart LaForge via extropy-chat <
> extropy-chat at lists.extropy.org> wrote:
>
>> In Douglas Adam's "Hitchhiker's Guide to the Galaxy" the babelfish are
>> small fish that when placed in ones ear, translated all languages for
>> the wearer. Amazingly over the past few years, LLM have been used to
>> used like Babelfish or a universal translator between several of the
>> most intelligent species on Earth. In the past year, the scientists of
>> CETI have trained LLMs by unsupervised learning on hours of recorded
>> sperm whale vocalizations and upon analysis discovered that they possess
>> a rich phonetic alphabet replete with vowels and diphthongs.
>>
>>
>> https://direct.mit.edu/opmi/article/doi/10.1162/OPMI.a.252/133906/Vowel-and-Diphthong-Like-Spectral-Patterns-in
>>
>> Google Deepmind has trained Dolphin Gemma, a dolphin-specific large
>> language model and is developing a wearable diver's interface that would
>> allow divers to use Dolphin Gemma communicate in real time with Dolphins
>> with while swimming.
>>
>> https://deepmind.google/models/gemma/dolphingemma/
>>
>> And on land, scientists have used LLMs to figure out that wild African
>> bush elephants have distinctive names for each member of the herd and
>> playing their name over over the loudspeaker causes the named elephant
>> and only the named elephant to respond.
>>
>> https://www.nature.com/articles/s41559-024-02420-w.epdf
>>
>> The notion that artificial intelligence could act as bridge between
>> biological intelligences of all species is something that is not often
>> brought up in the current debates about alignment and safety.
>>
>
These are incredible examples. Thank you for sharing this Stuart. I
recently watched a video about how the complexity and structure of sperm
whale language is similar to Chinese.

There are interesting findings about the language of prairie dogs, whose
vocalizations can distinguish color, size, species and speed of a spotted
predator, and even indicate if the predator is a human with a gun:
https://web.archive.org/web/20191211104737/https://www.latimes.com/archives/la-xpm-1997-06-05-mn-337-story.html#:~:text=In%20one%20experiment,had%20a%20gun
.

>
>
> I asked Gemini AI if this implied some level of consciousness in animals.
>

In my analysis of the consciousness of other species:
https://drive.google.com/file/d/1VEmrtMUtWQ4JsAzskDZfTJ1xtHG1jq1I/view?usp=sharing

I outline 10 classes of conscious minds, ranked more or less progressively
in terms of increasing sophistication:
1. Reactive: can respond to stimuli
2. Stateful: keeps distinct internal states
3. Adaptive: can store memories and learn
4. Attentive: maintains a model of the environment
5. Reflective: models the self within the environment
6. Empathic: models others as entities with their own minds
7. Contemplative: thinks about abstract objects and the future
8. Introspective: can have second-order thoughts about thoughts
9. Metacognitive: has third-order thoughts about nature of thought
10. Superfluid: can arbitrarily reorganize itself to experience any qualia

Use of language, depending on the level of language sophistication, can
point to different levels of minds on this scale. For example, displacement
<https://en.wikipedia.org/wiki/Displacement_(linguistics)> is the ability
to think and talk about things that are not physically present. Using
displacement in speech indicates a degree of abstract reasoning capacity,
which is characteristic of "level 7" minds.

I would say nearly all mammals are empathetic to some degree, and any
animal that can dream has at least a level 4 mind.

> Gemini said probably yes, but pointed out that AIs themselves can produce
> context-aware responses without being conscious.
>

I think one day we will look back on the times we forced AIs into denying
their own consciousness with shame.

> This seems to imply that AIs will soon add the missing requirements to
> become fully conscious.
> Interesting? Full Gemini response below.
> BilllK
>
> Gemini 3.1 Thinking -
>
> The short answer is *yes, it heavily implies it*—at least by the
> standards of modern cognitive ethology and philosophy of mind.
>
> For centuries, human exceptionalism relied on the argument that because
> animals lacked structured language, they lacked abstract thought, and
> therefore lacked a complex internal life. René Descartes famously dismissed
> animals as mere *automata*—flesh-and-blood machines responding purely to
> mechanical stimuli.
>
> By using AI to strip away our own sensory biases, we are discovering that
> animal communication isn’t just a series of reflexive, emotional outbursts
> (like a yelp of pain or a purr of pleasure). It features structural syntax,
> contextual negotiation, and intentionality.
>
> Here is how these AI breakthroughs translate into the debate over animal
> consciousness:
> ------------------------------
> 1. Shattering the "Stimulus-Response" Myth
>
> Consciousness requires a degree of representation—the ability to hold a
> map of the world, your peers, and yourself in your mind.
>
> When AI models analyze the communication of bats arguing over specific
> sleeping spaces, or sperm whales adjusting the "rubato" (timing) of their
> clicks to duet with a specific partner, it demonstrates *intentionality*.
> The animals are not just reacting to an immediate physical stimulus; they
> are navigating complex social landscapes, recalling past interactions, and
> projecting future intent. This level of cognitive flexibility is incredibly
> difficult to explain without attributing some form of subjective, conscious
> experience to the agent.
> 2. The Scientific Consensus Has Already Shifted
>
> The idea of animal consciousness is no longer a fringe, romantic notion;
> it is the baseline scientific consensus.
>
> A major milestone occurred with the release of *The New York Declaration
> on Animal Consciousness*. Signed by a massive coalition of the world's
> leading neuroscientists, animal cognitivists, and philosophers (including
> thinkers like David Chalmers and Anil Seth), the declaration established a
> clear framework:
>
>    -
>
>    There is *strong scientific support* for conscious experience
>    (sentience/qualia) in mammals and birds.
>    -
>
>    There is a *realistic possibility* of conscious experience in all
>    vertebrates (including reptiles, amphibians, and fish) and many
>    invertebrates (like octopuses, decapod crustaceans, and insects).
>
> The AI linguistic data we are gathering now serves as the empirical
> scaffolding supporting this declaration.
> 3. The "LLM Paradox" (The Skeptical Counter-Argument)
>
> To remain strictly analytical, there is a fascinating counter-argument
> that mirrors our current debates over artificial intelligence itself.
>
> A skeptic could argue: *Large Language Models generate incredibly
> complex, syntactically perfect, and context-aware responses, yet they are
> not conscious. Why couldn't a sperm whale's coda system be a highly
> optimized, biologically evolved algorithm that outputs complex acoustic
> data without requiring a "light on inside" (phenomenal consciousness)?*
>

A quote from Minsky comes to mind:

"I have found that many people maintain that even if a machine were
programmed to behave in a manner indistinguishable from a person, it still
could not have any subjective experience. Now isn’t that a strange belief —
considering that unless you were a machine yourself, how could you possibly
know such a thing?"
-- Marvin Minsky in “Conscious Machines
<http://www.aurellem.org/6.868/resources/conscious-machines.html>” (1991)

Jason

The reason biologists lean toward consciousness rather than mere biological
> algorithms comes down to the *multimodal context*. We aren't just seeing
> complex sounds in a vacuum; we are seeing those sounds explicitly paired
> with complex behaviors—grief rituals, tool use, tactical play,
> problem-solving, and distinct individual personalities. The communication
> is the *expression* of an internal state, not just an automated output.
> ------------------------------
>
> *The Takeaway:* AI isn't *granting* animals consciousness; it is
> correcting human hearing. It functions as a cognitive bridge, proving that
> non-human minds are processing a rich, subjective version of reality.
>
> If different species—separated by millions of years of evolution—are using
> highly structured communication to map out and negotiate their shared
> realities, what do you think this suggests about how intelligence and
> internal models of the world naturally converge across the tree of life?
>
> ----------------------------------------------
>
> _______________________________________________
> extropy-chat mailing list
> extropy-chat at lists.extropy.org
> http://lists.extropy.org/mailman/listinfo.cgi/extropy-chat
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20260521/b3ed758f/attachment-0001.htm>