Discover more from Life in the 21st Century
“Our conventional response to all media (technology), namely that it is how they are used that counts, is the numb stance of the technological idiot.” Marshall McLuhan, Understanding Media, 1964
The Artificial Intelligence (AI) hype-boom now in full flower can best be simply understood as the next generation of compute technology. From its beginning, we've had no real discussion about this technology, though why should that be any different from when our ancestors first flaked rocks to create cutting tools or picked up a good heavy stick to wield as a club? The entire history of technological innovation and implementation has occurred with little thought in regards to what the adoption of and then adaption to any specific technology means for the larger society. Today, we remain as transfixed by technology as those who long ago stared into the first campfire.
A flood of advertising now appears stamped with AI, the latest symbol of an ill-defined progress that most accurately represents the next step in brute force compute, allowing more powerful search capabilities and greater sophistication in compiling results. Only a few decades ago, similar fairy dust marketing was cast and problem solving promised by the internet's initial promoters. However, there is a difference between today’s AI hype and the initial net hype, this time it's not all rainbows and lollipops. Many of the first-in, those shouting full steam ahead, simultaneously warn of the technology's hazards, sort of like maniacs with loaded guns pleading not to let them shoot you.
Case in point is an article in the FT reviewing several books on the topic. One book is written by Mustafa Suleyman, who was at Google for a number years working on AI. Suleyman now has his own AI play, together with Reid Hoffman, who I previously noted issued his own please stop us from shooting you warning. Suleyman hypes any number of AI fantasies then warns, “Containment is not, on the face of it, possible. And yet for all our sakes, containment must be possible.” He writes, “I regard the often dismal picture painted in the following chapters as a titanic failure of technology and a failure of people like me who build it.” But continue building he will, with the irrational rationalization, “Make no mistake: standstill in itself spells disaster.”
Intelligence has always been a rather nebulous term. It's meaning argued over by philosophers, scientists, teachers, and parents. One thing for certain, Homo sapiens definition of intelligence is entirely prejudiced by our own experience. Webster defines sentient as “capable of sensing or feeling: conscious of or responsive to the sensations of seeing, hearing, feeling, tasting, or smelling.” However, we consider intelligence abstract thought, the ability to manipulate symbols, an activity beyond the senses. Nonetheless, the vast majority of the brain's activities deal with the functions of the senses and other operations of the body such as your heart beating and lungs breathing, this too is very much intelligence.
What we popularly conceive of as intelligence might better be understood as nothing more than an incidental result of the rest of the brain's functions, all innumerably repetitive activities, exactly opposite of the popular notion of great intelligence symbolized as “a stroke of genius.” Seventy five years ago, the importance of almost infinite repetition to the brain's activities allowed Norbert Wiener to understand developing information technologies would fairly rapidly allow automation of innumerable tasks. Today, the greatest power of this next generation of compute technology will be its ability to automate more industrial, agrarian, and information occupations. Not that this isn't a process already decades in motion, ask the UAW for example, but it will be of greater scale and breadth across society.
The questions we face aren't simply how to automate, but what might be automated, and most imperative, what does any specific automation mean for the structure of larger society? For this we need politics, unfortunately we are bereft of any politics of technology or any healthy politics for that matter. A politics for compute technology would consider two important factors: the design, the architecture of the machines themselves and the larger shaping forces the technology unleashes upon the greater society. The former factor impacts every compute process, the latter instrumental in defining society as a whole.
There are numerous components to consider in regards to the technology itself, most importantly, all output is completely reliant on the input. If the data input is prejudiced, insufficient, colored in anyway, so too will be the output. A machine can be locked into established ways just as easily a person. The architecture of the machine affects both what's inputted and the output, maybe with digital technologies in no greater way than reducing all data, all signals, to a binary code of either one or zero, an artificial dichotomy foreign to the complexities of even the simplest aspect of the biological world. This in no way denies the power of this process, but requires much greater understanding as it threatens to shape every aspect of society and life itself.
More important than the individual machines are the ever growing digital networks connecting the machines and processors of any given gadget, whether automobiles or door locks. This next generation of compute technology, AI, the brute processing force of billions of transistors is completely dependent on the continuous searching of network connected machines and devices for both old and new data. This data is compiled and used to respond to queries. In gaining specific results, these queries are completely dependent on repetition, “learning models.” The searching of data, the resulting compiling, and finally the output are all designed order. As with all designed order, whether the programs are created by humans or machines, they incorporate specific systems of power, power increasingly in the hands of fewer and fewer people. It is analogous to laws, which across most recorded history only a few had the power to initiate, change, and influence, except now, we in part begin to abdicate this power to machines.
The machines internal design and networking are only one side of the technology's political equation. The other, simultaneously influencing the machines design, are their interaction with whatever greater environment they are part. This is especially apparent in regards to automation. What might be automated is an essential political question, each line of code locks in a certain power, most assuredly power for some massive corporate entity.
The question of what should be automated is essential. Just as the current mad rush to electrify the established industrial economy as a supposed climate solution, best example here American car culture, there will be just as an insane push to automate the status quo of every present aspect of life without any discussion on how things might or should be done differently. Different outcomes would require not simply a technological perspective, but numerous perspectives of the greater society — political discussion.
Ironically, in many ways, just as with industrialization, the most imperative initial question being ignored is how this information technology fits best into the greater natural world. Industrialization changed the natural world through the brute force enabled by massive fossil fuel use. AI, ironically, requires brute force compute, a process, disregarding its specific outcomes, with massive impacts on the greater ecological systems they are part.
An essential recent article by the AP reveals the necessary unrestrained resource usage, particularly of energy and water, AI requires. AP writes,
“As they race to capitalize on a craze for generative AI, leading tech developers including Microsoft, OpenAI and Google have acknowledged that growing demand for their AI tools carries hefty costs, from expensive semiconductors to an increase in water consumption.... But they’re often secretive about the specifics.”
The article reports Microsoft's previous opening of massive data centers in Iowa for what's idiotically marketed as “the cloud”. Iowa has the essential water availability, for now anyway, necessary to keep the massively energy intensive data systems cool. It's sort of amusing with our growing concern, still not nearly enough, on industry’s impact on the environment, the tech-industry develops a new process with massive energy and resource needs, and the kicker, then try to market it as solution to the problems of massive resource use.
It's somewhat puzzling how information technologies have yet to greatly impact the structures of the two greatest political organizations of industrialism – the corporation and modern representative republicanism. After all, industrialism completely restructured the feudal political order of monarchy and the economic order of the large landed estate. Maybe it remains only a matter of time, though industrial corporations appear to be readily adapting to the new technologies, while most established government structures grow ever more dysfunctional, yet there remains no call and even less effort in thought or action to reforming either.
In Understanding Media, technology historian Marshall McLuhan writes, “For an increase of power or speed in any kind of grouping of any components whatever is itself a disruption that causes a change of organization.” This age of compute technology has seen a massive increase in the power of compiling information and ever increasing speed of communication, but presently led only to an entrenching of industrial organization, ever greater centralization of power in ever fewer massive corporations. A more beneficial utilization of these technologies in more substantive ways requires a much more radical restructuring, a replacement of established institutions and processes. Any such reorganization would certainly be met with great opposition by the established powers that be, yet the thinking and action required to even begin is lacking, the politics nonexistent. Such reorganization requires truly radical thinking about the organization of power, radical in the original Latin meaning of going “to the root.”
Across recorded human history, power has mostly been extremely centralized, hierarchies of order imposed from the top. The relatively short experiences with democracy in Ancient Greece and Rome, Renaissance Italy, and the last two centuries experiments with modern republicanism, provide our only understandings of more distributed, horizontally organized, political power. Such experience is extremely valuable and should in no way be discounted as it unfortunately has in recent decades. It needs to be combined with the knowledge of our contemporary technological environment and transcended to something new.
To this point, there are completely insufficient attempts at a politics of information technology through controlling “personal” information, an upside down, a doomed to failure politics as the labor/management dichotomy was for the politics of industrialism. The question isn't how the information is used, but what is the architecture of the system, the structural design of power. As labor, no matter how well organized, was always a servant to management in the established binary industrial order, whether capitalist or socialist, any individual node with whatever rights will be subservient to those in control of any centrally organized information network.
This next generation of compute technology is hawked as Artificial Intelligence. Intelligence is something we associate with the brain. The brain is a massively complex, distributedly ordered system. There are no CEO neurons or congresses of representative synapses. It is distributed order of a complexity and sophistication we still don't well understand. The centralized control of this technology will not only be politically destructive, but what the technologists don't yet understand, it will eventually prove fatal for the technology itself. If this artificial intelligence is to in anyway be useful in meeting the challenges humanity faces in the 21st century, it will need a lot more natural intelligence and more than a little, to this point completely lacking, wisdom.
Subscribe to Life in the 21st Century
History, Science, Energy, Technology, Environment, and Civilization
Have you seen this yet Joe?
https://retractionwatch.com/2023/09/26/nature-flags-doubts-over-google-ai-study-pulls-commentary/
Thought you would be interested..
"It's sort of amusing with our growing concern, still not nearly enough, on industry’s impact on the environment, the tech-industry develops a new process with massive energy and resource needs, and the kicker, then try to market it as solution to the problems of massive resource use."
Yup....and yet even in my head, it is not quite as instinctive or maybe I should rather say intuitive to associate those industries with ecological destruction quite as much as say agriculture, mining or fossil fuels...