The razor-sharp knife rests effortlessly in one hand, while the other orchestrates with poised assurance, steering clear of the unforgiving edge. The chef moves with liquid grace, with fluid and swift movements the ingredients yield to his expertise. Each gesture flows into the next, guided by intuition honed through countless repetitions. He knows what is necessary, how the ingredients will respond to his hand and which path to follow, but the process is never exactly the same, no dish is ever truly identical. While his technique is impeccable, minute variation and the pursuit of perfection are always in play. Here, in the subtle play of steel and flesh, a master chef crafts not just a dish, but art. We're witnessing an artisan at work. (This paragraph was co-authored by a human.)
Viewing AI through the lens of the artisan or craftsman reveals many intriguing questions. Who are the ‘master chefs’ of AI? At the core of generative AI are the adept engineers and AI researchers. These professionals possess the specialized knowledge required for coding and software development, acting as modern-day craftsmen who create intricate algorithms and foundational models. Over recent decades, this group has narrowed to a select few elite engineers and scientists, mostly working at powerful tech companies. On the other end of the spectrum is the general public, the majority of us, who merely use these algorithms, engaging with AI mostly through its interfaces. Rather than grasping the underlying principles, our aim is merely to ‘master’ the use of these tools and to seamlessly incorporate them into our (working) lives. Moreover, beyond this basic knowhow, today individuals may take it upon themselves to excel in 'prompt engineering', which at its core, involves mastering the art of writing effective prompts. But what is clever today can be obsolete tomorrow. Given the rapid progress in the field, what expectations should be set for the wider public and workforce regarding the cultivation of new AI competencies?
Our first interactions with technology are typically centered on the physical tech devices and technological systems we use every day. Yet, the terms 'technics' or 'technique' encapsulate both the notion of technical products and technical activity. While the terms artifacts and tools denote entities of technology, the Greek origin of the term, technè, signifies craftsmanship or skillful know-how.
This original meaning aligns with what we currently refer to as 'practical knowledge’. However, we have to be careful with this translation. In some discourses, there exists a rigid demarcation between the theoretical, objective knowledge found in science and hands-on skills such as carpentry and cooking. True knowledge can only be found in the first domain. In the second domain there is only the practical application of knowledge. Whereas theoretical knowledge aims at uncovering truth and understanding reality, practical knowledge focuses on making things and changing reality. Although these distinctions can be traced back to Aristotle, his original differences between theoretical and practical knowledge were more subtle. For Aristotle, technè not only referred to the carpenter but also to the doctor, who ‘produces’ health through therapies. Both types of knowledge pertain to the understanding of knowledge as a capability that encompasses active and passive components. Thus, theory extends beyond mere contemplation, and practical knowledge transcends simple production.
Furthermore, recent decades have seen a consistent rise in research focused on embodied knowledge. As implied by the term ‘know-how’, the practices of hands-on skills are fundamentally grounded in knowledge as well. For instance, a sportsman like Messi harbors an innate understanding of soccer dynamics, he knows how to take a freekick, something Michael Polanyi labeled tacit knowledge. Likewise, the carpenter knows how a type of wood will respond to his actions and what the best way is to treat the wood for certain purposes.
All skills require (tacit) knowledge. Similarly, all scientific knowledge requires skills. Even seemingly abstract fields like mathematics demand a certain practical skill set, with practitioners wielding tools and methodologies to unravel complex problems, such as using pen and paper in algebraic operations. This is perceptibly mirrored in the Dutch language, where terms such as ‘wiskunde’ (mathematics) and ‘geneeskunde’ (medicine) hint at this shared ground of ‘kunde’ (technè) in theoretical and practical knowledge.
Yet, while overlapping in their reliance on skillful know-how, the type of knowledge of the mathematician also diverges from the typical crafts. A mathematician brings forth eternal truths, aiming to derive conclusions anchored in foundational axioms, venturing in the realm of the Absolute. His skill enables him to solve problems ‘actively’ on pen and paper, as he works step-by-step, following certain fixed rules and procedures. Body and mind work together here. I always like to think about it as surgery on paper, making incisions using a pen or chalk. If you make one little mistake, the entire operation is screwed.
Conversely, the craftsman, the physician, or the athlete navigate the contingent and variable phenomena of daily life, wherein knowledge is not static and universal but dynamic and particular. It adapts to the changing circumstances, grounded in the practical and the experienced. This nuanced differentiation and similarity between ‘technè’ and ‘epistème’ — practical and theoretical knowledge — is a cornerstone of Greek philosophical thought and still very influential today, for example in our thinking about the entangled nature of science and art in the ‘production’ of facts. We do not only discover facts, we built them.
Likewise, there is no life without technics and no technics without life. This perspective shines notably in our interactions with everyday tools and machinery, when one shows a certain handiness in using hammers or computers for example. We simply couldn’t live without these devices. However, more fundamental, it underscores that even when engaged in activities perceived as intrinsically human — such as brainstorming or creative thinking — there exists a foundational layer of technè or skillfulness. For example, consider a masterful storyteller, who employs a rich array of techniques to convey a persuasive and compelling narrative. Sometimes they are born speakers, but often their prowess is not simply a birthright but a cultivated skill, a technè honed over time. You need to practice to become a master in writing or speaking. Life thus also means becoming technical, in the sense of learning all these different skills and know-hows through the course of your life (Susanna Lindberg, 2023). Hence, the art of living.
While the great poet becomes a master of language techniques, using his tongue-as-a-tool, most skills point to the extension of biological organs through external instruments. Traditionally, the skillful person has been linked with these tangible, physical tools that become extensions of ourselves, seamlessly integrated into our bodily experience, such as riding a bike. Yet, in today's digital landscape, the emphasis has pivoted towards the cultivation of digital proficiency. Today, policy makers often stress that we need to develop ‘digital skills.’ In today's app-driven economy, which is dominated by as-a-service business models, it isn't necessary anymore how to repair your bike or understand its mechanics (under-the-hood). Instead, it's essential to know how to use ‘the app’ to summon a mechanic.
But what exactly constitutes the skill set required for app proficiency? Amidst the layered landscape of digital systems, many find their expertise confined to surface interactions, predominantly at the interface level, lacking a deeper understanding of the underlying principles governing the digital world. This presents several risks to society. Borgmann's (1984) classic analysis reveals how this ‘apparatus paradigm’ can lead to an estranged relationship with the material world, where we become detached from the workings and craftsmanship behind everyday objects. This detachment, in turn, can blind us to the inner workings of our tools, making it easier for external forces to guide our actions and decisions. Without a deep understanding of the devices and services we use, it becomes challenging to form a healthy and meaningful connection with them. Borgmann cautions against such a disconnection, suggesting that it diminishes our engagement with the world around us.
Consequently, a growing group of users and policy makers long for a deeper connection and understanding of the foundational layers of the digital environment. They want to bridge the gap between surface engagement and profound expertise. Amidst the turbulence of online environments, many argue it is becoming increasingly imperative for citizens to harbor a more technical understanding of digital realms. We need to pursue and cultivate skills that transcend mere app proficiency and improve our digital resilience. This encompasses interface skills such as discerning credible information online and lower-level skills like programming capabilities. Accordingly, this skill set encompasses not just the ingenuity in tweaking apps but also a critical approach to navigating digital environments and recognizing how algorithms influence behavior.
When we now introduce AI into this discussion, it invites scrutiny of the competencies that are valued and necessary within this domain. Beyond the highly specialized but very small population of super engineers and AI scientists, what should be expected from the wider populace? This includes the young and the elderly, individuals with high levels of education as well as those with less, digital natives alongside newcomers to technology, and both students and working professionals. One way or another, they must all cultivate digital AI proficiency to establish a healthy online environment.
However, this ambition can easily turn into a hollow ideal. What AI skills should we focus on developing, and who is the target audience for these skills? For instance, is there a genuine, widespread need for basic education in coding and AI literacy for everyone? Although having a basic understanding is certainly beneficial, it prompts a deeper inquiry into its practical value against the backdrop of complex Large Language Models (LLMs) and sophisticated digital infrastructures. The necessity for a common man to delve into the basics of coding remains debatable. Overwhelmed by the current wave of complexity from foundational models, do we really win something with some very basic programming skills and knowledge of neural nets for most of us? I hardly think so. The nature of digital systems makes ‘under the hood’ knowledge unrealistic. Furthermore, many individuals have developed a fairly intuitive understanding of how algorithms operate, for instance on social media platforms. Yet, what impact does this knowledge have?
As part of a broader strategy, improving digital literacy about the lower layers of the stack can hardly be argued against. Learning how to write code (fundamental layer) and build or understand algorithms could help us better understand what happens at the interface level, but people will be mainly focused on developing skills at this interface layer. The daily interaction with AI for most individuals happens not through coding but through various software and interfaces layered over digital platforms. Indeed, the rise of generative AI has spotlighted prompt engineering as a highly demanded new skill in the corporate sector, eclipsing the need for coding expertise. It encourages individuals to focus on honing their ability to prompt effectively rather than getting tangled in the complexities of coding or understanding AI models.
However, this approach is not without its challenges, as the superficial nature of this skillful know-how can leave us vulnerable and exposed, as Borgmann explained. Additionally, we find ourselves in a perpetual cycle of adjusting to new interfaces, with a relentless progression of new skills to learn, new apps to master, and fresh software suites to familiarize ourselves with. For instance, in the case of prompt engineering, envisioning a 2D text box as the pinnacle of generative AI interfaces seems rather limiting. Are we really going to interact with AI models through open-end prompting? The future arguably holds a multitude of AI tools and interfaces necessitating diverse skill sets, guided by the specific tasks at hand and the integrated software solutions. Nonetheless, this poses challenges in developing a skillset strategy. Companies or policymakers might concentrate on honing a particular set of AI skills for citizens or employees, only to find those skills outdated in a few years. Alternatively, opting for a basic understanding for everyone might seem prudent, yet the lack of specialization raises doubts about the actual benefits or advantages of acquiring this basic set of skills.
Take, for instance, the generation of images through AI; while some find it exhilarating to conjure images through simple prompts, super easy for all of us, others with a background in visual arts or proficiency in tools like Adobe Photoshop may prefer a seamless AI integration within such advanced platforms. Similarly, the way writers leverage AI tools in the future, could range from utilizing simple queries on interfaces such as Bard or ChatGPT for drafting initial social media posts, to anticipating the integration of LLMs in mainstream text editors like Microsoft Word or even operating software.
Additionally, our interactions with AI are not solely direct; often, they occur alongside other (technical) activities. Just as other digital tools did previously, AI can augment or supplant existing skills, evident in activities like writing or photography. Consequently, the rise of generative AI does not only demand a process of ‘reskilling’ but will also lead to an interplay of ‘upskilling’ and ‘deskilling’. For instance, it's pivotal to acknowledge the already subtle yet pervasive presence of AI in daily digital experiences, be it autocorrect features refining our texts or the optimization of digital photographs. The implications here are twofold. First, the skill of texting or taking a photo on your smartphone is not so much directed towards AI but affected by it nevertheless. Second, AI has profoundly altered these forms of art and skills as well. This delineates a paradigm shift in skill mastery. The art of photography, for instance, has transitioned from manipulating the settings of a traditional analog camera to adeptly using a smartphone, where algorithms work tirelessly on the background to facilitate this newfound form of digital craftsmanship. This may not apply to the professional photographer, but it certainly holds true for the layman.
Building upon this notion, it is evident that the recent surge of generative AI is not only reshaping but also recalibrating our perception of true art and creativity, elevating some crafts over others. The discourse surrounding the true essence of an artisan gains traction as the boundaries between human and machine-driven creativity and production become increasingly blurred. For instance, the advent of prompt engineering as a fresh craft has sparked dialogues regarding its potential to forge unprecedented forms of art. Are we on the brink of recognizing a novel category of artisans specialized in ChatGPT and Midjourney creations? Or is this nothing else than the final blow to what we once named creativity?
Historically, we've engaged in similar debates, likened to those contrasting photography with painting, and cinema with theater. Each transition entails a social endeavor to discern which elements in the evolving processes of creation—where life and technics are invariably intertwined—merit the designation of craftsmanship, and which outcomes deserve safeguarding through intellectual property laws, distinguishing them from those that do not.
As evidenced by the ongoing disputes in Hollywood and legal battles involving current generative AI enterprises, navigating this emerging landscape promises to be a turbulent journey.
In conclusion, cultivating a wide range of AI skills is essential. Yet, the extent to which the average person needs to delve into coding, programming, understanding AI or simply cultivate interface skills remains a complex issue. We cannot expect too much from skills below the interface layer, however too much blind focus here makes us vulnerable. It demands a nuanced balance between maintaining independence and depending on the ever-advancing AI tools and software in the years to come. This means a balance between embracing generative AI, but also the capacity to say no and prioritize working on other skills. Similarly, the adoption of 'AI creations' as widely recognized forms of art is unlikely to occur instantly but will likely unfold through a challenging process of social negotiation with technology. Furthermore, the conversation about digital competencies should be integrated into the broader discourse on fundamental skills and knowledge, like literacy, numeracy and critical thinking. The impact of AI is not limited to merely requiring new skills; it also involves a process of technological mediation. This means that generative AI shapes our thought processes, communication, reasoning, and writing abilities.
(This article was co-authored by AI.)
