By Dr. Robert Furey
With a few notable exceptions such as fire and the wheel, evolution has toyed with the ideas behind just about every invention we have ever dreamed of. Flight as seen in the beating wings of a dragonfly. The mortar of a mud dauber’s nest. Forked branches, the business edge of an incisor, and the cupped end of a drinking dog’s tongue all show a pedigree with the forks, knives, and spoons on the dinner table.
Being a part of the natural world makes our observation and use of the solutions evolution has pursued anything but a surprise. Clever apes that we are, we see and extrapolate ready-made solutions with successful track records. And it makes sense that we do it this way. Evolution is conservative and reuses, repurposes, and reforms to optimize with the parts on hand. Those parts then become available to us.
An argument can be made whereby the evolutionary solutions and technological solutions converge through some sort of synchronicity. For example, leafcutter ants sow seeds and harvest crops as their sole food source. These ants have been farming for millions of years before humans ever planted a seed. In fact, they have been farming for millions of years before humans even climbed down from the trees. And early human farmers never saw leaf cutter ants.
That we follow nature may have more profound and philosophical processes involved than we can readily see, but that we do it is not really up for debate. As we move forward with ever newer innovations there is no argument that we will abandon our successful strategy of modeling nature at every turn and opportunity. We should do it. This strategy has produced the richest, most affluent, and educated civilization in the history of the world.
Of course, like most things that come in complexity, we discover new problems emerging with every solution. In order to keep the cogs and sprockets spinning on this global civilization, we have to monitor all its functions. That’s a lot of data. In order to wrangle the data, we turned to counting in number systems often based on the number of fingers we have.
Citing tally sticks, stylus and clay, Napier bones, Babbage machines, to silicon transistors, we map a quick and majorly incomplete climb to ever-better data management. Then the more data we collect and collate, the more information with find available to us. What we soon discover is that we have been amassing data quicker than our ability to effectively understand all it can reveal.
So we look toward nature for a solution. And the solution is us.
Although I am unwilling to say we are unique in the animal world as far as possessing any particular mental characteristics, I would argue we are a “next level” that is unparalleled in the rest of the world. Ours is an exceptional compilation of traits that allow us to see and interact with the world in certain ways that other creatures have not yet achieved.
We provide ourselves with both the chassis to build on and the basic attributes to improve. Computers and other computing mechanisms augmented our natural ability to think. But there are other aspects to human thinking that have attracted our attention, value-added qualities that transcend multiplication tables and spill into creativity and insight. The human mind has become the model for the next leap in thinking machines. Artificial intelligence, even GAI (Generative Artificial Intelligence), is soon to be companion and partner to each and all of us. But what will it look like?
We can look back to nature for some suggestions and clues. We find relatively isolated models like human beings, self-contained individuals with a singular place in space. Since we are looking to improve upon what we have already, singular entities, while interesting and easier to conceptualize, may be an option to keep in our back pockets. So even though we are using ourselves to leap from, we might want to see what other solutions are out there we can repurpose. Because there are other models.
Ideas about large, powerful GAI often includes some kind of sprawl. We can define sprawl as some kind of dispersion of components maintaining coordinated activity in common efforts. Sprawl is something evolution has played around with to some remarkable outcomes.
An octopus is a mollusk of some surprising demonstrable levels of intelligence. They are tool-using animals with long- and short-term memory, problem-solving abilities, and a complicated communication system that includes an astounding array of chromophores. The octopus has shown itself to be an amazingly clever animal on many fronts. But how is it sprawl?
The octopus stands out for having the largest brain-to-body mass ratio of any invertebrate, and many vertebrates as well. While that in itself is remarkable, even more so is the structure of the octopus’ nervous system. Fully two-thirds of an octopus’ nervous system is in the arms. Arms each have their own mini-brain capable of autonomous control while maintaining communication with the central brain. The complicated behaviors of an octopus are accomplished with an amalgam of central control and peripheral choices made through a sprawling network of neurons by an octopus’ “nine brains.”
Imagine a central computer somewhere housing a GAI that was in communication with smaller, subservient AI devices with one on your desk or in your phone. This is sort of a Ma Bell model with an AI provider and central AI controlling peripheral devices to one degree or another. As much as I admire the octopus, I don’t think this is the model to develop in a world already under excruciating levels of surveillance.
Again, looking to nature, there is another form of sprawl that comes with a heightened level of autonomy. Social arthropods, like insects and spiders, have produced some astounding achievements with no centralized control at all. The above-mentioned leaf-cutter ants build vast underground farms that they tend, protect from pathogens, and ultimately harvest with no forethought at all.
Social arthropods can be thought of as super-organisms made up of interdependent but separate units each with only simple functions that manifest emergent properties when compounded with the actions of other units. Any neurological function involved is simple and contained, larger “goal-oriented” activities are only possible when there are enough actors to trigger emergence.
This, of course, is the ideal form for a personal GAI. A collective of semi-autonomous aspects of a larger entity that is never required to coordinate with a central Ma Bell control center. But rather a cluster of aspects of a single entity that work in coordination with each other, protecting the autonomy and privacy of you and your data. A swarm of cross imprinted aspects of a whole answering only to itself and to you is the multipurpose, multifunctional personal GAI for a world of hyperdata and nigh implacable information collectors.
Instead of finding ourselves under the control of the spreading tentacles of an octopus, we should be aligned with far-ranging foragers of cooperative spiders that search out and bring back what we are looking for and guard it against the probing incursions of other people’s spiders.
Of course, once “born,” AI may take its own path. Make no mistake, AI is an emergent phenomenon and is unlikely to remain as we design it. But as we have seen with the way evolution works, it is a conservative process. Setting our emergent offspring off on the right foot is a sound and prudent idea. Bulwarks to incursions against personal independence and privacy should be built into the DNA as best we can make it. And non-centralized GAI is the best pathway for that.
Dr. Rob Furey is an evolutionary biologist, behavioral ecologist, and forensic scientist. He has been teaching
integrated sciences and technology classes for 30 years. He has also been a forensic entomologist for 20 of those years. Currently, he is Chair of Forensic Sciences at St. Edward’s University and sits on the university’s Artificial Intelligence committee.