One Tool to Rule Them All

You’ve all heard it now a million times. Anytime anyone has the temerity to dare a conversation about AI, or smartphones, or the internet, someone in the group can be relied upon to intone sagely, “Well you know, it’s important to remember that technology is just a neutral tool; what matters is how you use it.” Heads will nod solemnly, you will stammer and stutter for a moment, then shrug and let it go, and another little opportunity to grow in wisdom together will have been lost beyond recall.

To be sure, there is a grain of truth in this cliche. Inasmuch as technology is designed as an instrument of human purposes, it remains responsive to the human will, which is almost infinitely varied in its desires, and whenever the technology is used, we attribute moral agency and responsibility to the user, not the tool. I may pick up a hammer to repair a neighbor’s house damaged in a storm, or to crack my neighbor’s skull in a fit of rage. If I do the latter, the law will take me, not Black and Decker, to task.

And yet as soon as we scratch beneath the surface, we find the concept of the “neutral tool” beginning to break down. After all, even the simple hammer has a lot to say about how we can and can’t use it. It may be used for either good or evil, but only within a fairly narrow range of applications—specifically, those that involve whacking things, and indeed only a small subset of those. It works just fine on a nail, but will wreck a screw. It might—especially in the form of a large mallet—work together with a wedge to split a log, but used on its own it will make almost no impression on the wood. Indeed, to speak of a hammer as a tool is precisely to highlight its limits, for a tool is, almost by definition, devised for a purpose, or a discrete range of purposes. It has, to use the Aristotelian term, a “final cause”—an end for which it was devised—and it is this that determines its “formal cause”—the structure in which it was designed. Every tool is built with certain ends or objects in mind, and generally works best when applied to those objects; applied to others, it may not work at all, or worse, inflict active harm. To be sure, every tool will have unanticipated uses, waiting for enterprising users to discover or invent; WD-40 was originally developed to prevent corrosions on missile skins. But even if we cannot anticipate them in advance, in retrospect they will usually bear some family resemblance to the primary ends for which the tool was designed.

This observation should prompt us to ask, “for what ends were our current digital technologies devised?” Of course, there are no easy answers; for technologies as complex as these, the causes and motivations behind their construction must be equally complex. We may point to the long and ever-present influence of the “military-industrial complex,” whose research priorities have shaped nearly all modern consumer technologies, from the face-recognition algorithms that help power Instagram to the GPS that guides your morning commute. We may point to the powerful incentives of profit and power, which helped drive the “surveillance capitalism” model of Silicon Valley to find ways to capture and harvest human attention. But perhaps the most striking feature of the modern innovation ecosystem is the extent to which it defies such final-cause analysis. 

For many of today’s cutting-edge innovators, the innovation has become an end in itself. J. Robert Oppenheimer, asked why he helped build the Bomb, is said to have answered, “When you see something that is technically sweet, you do it.” And such is the spirit of Silicon Valley that has given us everything from Uber to Tinder. Alongside apps devised as genuine problem-solving tools, our app stores are also full of applications that began as zany ideas—“hey, wouldn’t it be cool if…?”—which were then retrofitted with a monetization strategy. Rather than being built as tools to serve people and their concrete needs, our technologies today are too often forged as mere demonstrations of technique, which people are then enlisted to serve, in order to justify their continued development and deployment. Last year, I recall asking a leading engineer at a major tech company why they had simply added their AI companion to all of our computers and programs without asking users whether they wanted to use it. He was confused by the question: “Well, we just figured people would.”

Unlike the technological tools of yore, which were hung on a hook or stored in a toolbox until asked for, our new technologies are always with us, in our pockets and at our fingertips, constantly demanding to be deployed. This is because, to the extent that our most powerful tools today have been designed for a purpose, it is for the paradoxical purpose of being “general-purpose technologies.” Rather than being purpose-built like a hammer or chainsaw, devised and designed in service of some ends rather than others, technologies like the internet, the smartphone, or artificial intelligence are meant to be the ultimate Swiss Army knives, capable of being put to almost any end that the user desires. That is meant to be the beauty of them—as the advertisements enthuse, “the possibilities are endless!” This end-less-ness of modern technology reflects the chief end of man as re-imagined by our postmodern politics: a vision of freedom as a choose-your-own adventure story. Forge your own path, make your own way, follow your own dream—these were the mantras of the 80s and 90s, within which the internet came of age. In other words, if a technology always inscribes a certain purpose in its design principles, the purpose of our digital world was to conform itself to your desires, whatever they might be. 

Is it not then the case with digital technology that we have at last made the idea of the “neutral tool” a reality? The hammer, the drill, the sword are all finite, limited to the particular purposes for which they were built, solving the particular problems that bedeviled our ancestors. But by encoding the entire world into bytes and bits, and placing a universal decoder in each of our hands, have we not at last created a tool that is as infinite as the human mind itself—or indeed, with the creation of artificial superintelligence, far more so? Surely this super-tool, at least, is “neutral,” in the sense of being able to be applied to any object, shaped to any purpose that we might devise?

To ask this question is to answer it—at least for any attentive reader of J.R.R. Tolkien’s fantasy classic The Lord of the Rings. Although set in a pre-technological mythical world of wizards and dragons, Tolkien’s trilogy is, more than anything else, a profound meditation upon technology. Nor should this surprise us; for as C.S. Lewis has argued, the modern enterprise of “technology” and the medieval pursuit of “magic” are two versions of the same quest: to manipulate the world through the discovery of hidden knowledge, the perfection of arcane techniques. In Tolkien’s tale there are two types of “magic”: one represents a kind of creative creaturely participation in the supernatural that depends on loving attention to the shape of nature, while the other serves as a metaphor for modern technology as the quest for mastery by reshaping nature to our wills. It is represented chiefly by the “Rings of Power” and the palantiri or “Seeing Stones.”

Each are startlingly apt metaphors for digital technology, considering that Tolkien probably knew nothing of the work his fellow Englishman Alan Turing was then doing down the road at Bletchley Park to lay the foundations for the computing revolution. The Rings are a technology for exercising control at a distance, the Stones for gaining knowledge at a distance: “to see far off, and to converse in thought with one another,” as Tolkien describes it. The Ring tempts its wearer with “incognito mode,” the power of invisibility, but its end effect is the erasure of agency and personhood: “if he often uses the Ring to make himself invisible, he fades: he becomes in the end invisible permanently.” The Ring exercises a compulsive power upon its possessor to use it without even thinking: each of us today can sympathize with Frodo: “almost before he realized what he was doing, his hand was groping in his pocket.” And as for the palantiri, today after two decades of social media, I find it much easier than I did as a teen to understand the madness of Denethor, doomscrolling in his lonely tower for hours and days on end: “The knowledge which he obtained was, doubtless, often of service to him; yet … [it] fed the despair of his heart until it overthrew his mind.”

Whole essays could be written on what each of these technologies have to teach us about the age of AI, but let me highlight just two key points here, which expose the limits of the “tool” metaphor.

First, Tolkien understood that powerful technologies do not exist in a vacuum, each ready to hand for an individual user to operate at his own whims. Rather, they exist within networks, such that a new node in a network can retroactively alter the meaning and function of a technological tool. On Tolkien’s telling, the “three rings” were originally forged for the elven-kings to work the good kind of magic: to create things of beauty and wonder and to heal and nourish. But once Sauron forged the master Ring, more powerful than the three, neither they nor the other rings of power could be used without falling under the power of Sauron’s malign will, which sought to draw all things under his control. Indeed, the rings were so interconnected that, possessing the One Ring, Sauron could read the thoughts and see the deeds of all the other ring-wearers—a chilling thought that has become all too familiar to each of us in the days of algorithmic surveillance. 

Similarly, the palantiri, or Seeing Stones, are each interconnected or networked, although some are more closely linked to each other than others. While in principle each palantir, like a powerful search engine, allows its user to turn his gaze wherever he wishes in Middle-Earth, past or present, and even to dimly discern possible future events, in practice the palantiri too are subject to the superior wills of other operators. Thus, when Saruman looks into the palantir of Orthanc to try to gain more knowledge for opposing Mordor, his gaze is instead soon captured by Sauron, possessor of the stone of Minas Ithil. Sauron uses the two-way connection to directly manipulate Saruman, like a mysterious stranger interloping in your DMs, until Saruman slowly becomes little more than a puppet of Sauron’s purposes. For Denethor, the mechanism of corruption is slightly different; instead of speaking to him directly, Sauron simply “curates his feed,” we might say, showing him less and less of the things Denethor is trying to see, and more and more of those that Sauron desires him to see. The result is indeed that Denethor becomes addicted to “doomscrolling,” and ends his life in suicide, a startlingly prescient account of how our media feeds can shape us. 

Second, Tolkien understood that the more powerful the tool, the less likely that an ordinary user could use it effectively, rather than being shaped or controlled by it. Thus, when Pippin looks into the palantir of Orthanc, he is almost immediately captivated by it and wholly under the domination of Sauron’s will operating through it. Only Aragorn, the rightful possessor of the stone and a man hardened by battle and seasoned with wisdom, is able to “wrench the stone to his own will,” and that only through “a bitter struggle.” Similarly, a recurrent theme in discussions of the Ring is that it gives power to the wearer “in proportion to his stature.” Pathetic little Gollum is unable to make much more use of this most powerful of all earthly creations than to enable him to catch fish or steal more easily; the Ring degrades him rather than expanding his powers. Only those already very strong can bend it to their will, but even they should dare not, knowing that it will bend them to its: “Its strength,” declares Elrond to Boromir, “is too great for anyone to wield at will, save only those who have already a great power of their own. But for them it holds an even deadlier peril. The very desire of it corrupts the heart.” Thus both Gandalf and Galadriel, each powerful enough to bear one of the three rings, refuse the gift of the One Ring when it is offered them, fearing the tyrant that each might become.

To be sure, the Ring, because it was made by Sauron, was not a “neutral tool” even in the beginning; it is bent towards evil, so that no good intentions can ever finally master it. But the palantiri can be used for good, as information technologies for the pursuit of truth and beauty—or, as Aragorn ultimately uses the stone, to deceive and defeat the wicked. Even so, Tolkien warns us to see their apparent “neutrality” as a powerful temptation: speaking of the palantir, Gandalf says, “And how it draws one to itself! Have I not felt it? Even now my heart desires to test my will upon it, to see if I could not wrench it from him and turn it where I would—to look across the wide seas of water and of time to Tirion the Fair, and perceive the unimaginable hand and mind of Feanor at their work, while both the White Tree and the Golden were in flower!”

Returning from Middle-Earth to our own time, what does this excursion into myth have to teach us about the myth of technological neutrality? Quite simply: the more powerful the tool and the more “general-purpose,” the less likely that we will be able to effectively master it and subordinate it to our purposes. We know this well enough from everyday experience: even a child can soon learn to use a hammer well, but the same cannot be said for a chainsaw, or an automobile. Part of this is due to the crucial distinction between tools and machines, too often ignored in this context. In a key passage of The Human Condition, Hannah Arendt writes of this distinction:

“Unlike the tools of workmanship, which at every given moment in the work process remain the servants of the hand, the machines demand that the laborer serve them, that he adjust the natural rhythm of his body to their mechanical movement….Even the most refined tool remains a servant, unable to guide or to replace the hand. Even the most primitive machine guides the body’s labor and eventually replaces it altogether.” 

With the stage that automation had reached when she wrote in 1958, she observed that we may well wonder “whether we are the masters or the slaves of our machines.” 

Today, as we have moved from tools and machines of the world to tools and machines of the mind, this worry is only more urgent. For as long as we controlled mere dumb machines with the tap of a button or the swivel of a joystick, the fact that they had far outgrown the power of our bodies need not unduly trouble us; man remained the master through the power of his mind. But once we immerse ourselves in technologies of the mind (like the Rings or the Stones) and move to create thinking machines, we are in grave danger of adapting even our highest faculty to the rhythm of our “tools.” Saruman, the ultimate technologist in the trilogy, “has a mind of metal and wheels,” laments Treebeard, “and he does not care for growing things, except as far as they serve him for the moment.” He views the world, as modern man has come to, in purely instrumental terms, with nature as mere raw material to be put toward some new and higher purpose.  

But as we become conformed to the image of the machines we built, as Tolkien understood, we instrumentalize ourselves, engaging in what Byung Chul Han calls “auto-exploitation,” ceasing to be individuals and becoming mere nodes in a network of other corrupted human minds. The result is that these technologies, networked together from mind to mind, fall under the dominion of only the strongest minds, those capable of bending them—and all of their other users—to their will. As Lewis warned in his Abolition of Man, “What we call Man’s power over Nature turns out to be a power exercised by some men over other men with Nature as its instrument.”

If we meddle carelessly with tools more powerful than us, in short, we will find ourselves reduced to mere tools, slaves of our own creations or of those powerful enough to use them against us. The grasp for powers beyond our reach ends in the destruction of all our native powers, the erasure of our agency: in Denethor or Gollum’s self-immolation, in the eternal fading of the Ringwraiths, in the pathetic end of the great wizard Saruman:

“To the dismay of those that stood by, about the body of Saruman a grey mist gathered, and rising slowly to a great height like smoke from a fire, as a pale shrouded figure it loomed over the Hill. For a moment it wavered, looking to the West; but out of the West came a cold wind, and it bent away, and with a sigh dissolved into nothing.”