The ubiquity of data coupled with the resultant proliferation of data mining-driven knowledge creation initiatives are reshaping the longstanding conception of what it means to learn. One of the core manifestations of the otherwise imperceptible shift is an explicit differentiation of reason-based and technology-based learning modalities. The former, which manifests itself as either direct observation-based experiential inquiry, or as scientific investigation-rooted theoretical investigation, encapsulates the familiar, intuitively obvious human-centric framing of learning. The latter, on the other hand, can be seen as a product of info-technological advances; as such, it can take the form of either computational, which entails algorithmic analyses of available data commonly characterized as ‘machine learning’, or the still developing simulational learning, where data and computing technologies are used to render physically inaccessible depictions of reality.
Implied in the so-broadened conception of learning is that learners should not be assumed to be human. That is because as artificial systems’ abilities to generate and ‘consume’, in the sense of taking learning-driven actions, information is quickly approaching – and in some contexts even surpassing – human sensemaking. Banks, insurance companies, retailers and a growing array of other volume-based business organizations now commonly rely on artificial learning systems to make eligibility determinations and identify potential fraud, and not only process but also to initiate transactions (the latter perhaps best illustrated by the now ubiqutuous automated trading). Interestingly, much of the knowledge generated by auto-processing systems is entirely contained in those systems, hence it is important to think of those systems as progressively more autonomous learning entities that are increasingly more capable of drawing inferences from data that go beyond the basic trend identification and extrapolation. And while the numerous and profound technological and moral implications continue to stir vigorous scientific and ethical debates, one aspect of the learning related impact of the currently underway info-technological transformation that received comparatively little attention is its impact on another longstanding notion: our conception of creativity.
Although learning and creativity have been traditionally seen as notionally distinct and different, the wave of rapid techno-informational transformations that sweep across an ever-wider array of commercial activities is strongly suggestive of the need for a reassessment of that perspective. To that end, the longstanding conception of creativity frames that somewhat vague idea in the context of imagination, commonly described as forming of new ideas, not already present to or in the senses. Moreover, the current conception of creativity also effectively limits it to the uniquely human, conjured-by-the-mind processes. Such characterization of creativity begs two fundamental questions: Firstly, should imagination-framed creativity be characterized as conjuring up of something completely otherworldly, in the sense of not having been experienced, or is it more appropriate to see it as forming something new – be it an abstract idea or a work of art – by combining ‘old’ elements into a new whole? And secondly, should creativity continue to be seen as a uniquely human, conjured-by-the-mind process, or should its definitional scope be expanded to encompass the more broadly framed capability of generating novel rearrangements of known elements?
While such definitional questions rarely lend themselves to a conclusive determination, evidence from the field of child psychology, which focuses on cognitive and behavioral human development, suggests that human creativity is closely linked to tacit and experiential learning. And if creativity is to no longer be limited to just ideas conjured by human mind, it seems reasonable to assert that artificial learning systems, which first identify patterns in data and then generalize beyond those patterns, should also be deemed capable of ‘imagining’. Those learning-capable computing systems – commonly grouped under the general label of ‘machine learning’ – can render not-yet-experienced (by human learners) ‘reality’, immersion in which could spark ideas – in human mind – that go beyond those that could be conjured by mind alone. In short, data, or more specifically data-enabled automation, should be seen not only as a source of numerous operational efficiencies, but also as a driver of creativity.
Mind-Machine Collaboration
Thinking about the prospect of computer simulation-enabled creative problem solving it is easy to get excited about the possibility of using those powerful capabilities to shed light on some of the most existential questions of mankind, but should business and other non-scientific organizations be as excited, given the comparatively mundane nature of their informational needs? The answer is a resounding ‘yes’!
Commercial organizations that seek to grow and prosper, and social organizations that aim to further different social and societal goals, pursue their goals in competitive environments, thus their ability to advance their agendas is tied directly to the efficacy of their functioning, which includes the use of already in hand or potentially available informational resources. Business companies can win more customers, and nonprofit organizations can secure the necessary resources by thoughtfully structuring organizational tasks in a way that combines the core advantages of informational automation – speed, scalability, and consistency – with core human competencies, most notably adaptively creative problem-solving. As illustrated by examples ranging from relatively simple tasks such as automated customer loyalty classification to much more complex ones, perhaps best embodies by self-driving cars, modern data analytics-driven systems can augment, or even replace direct human effort, which has the obvious benefit of freeing human time to be redirected toward addressing other problems. But all that is old news – modern, well-functioning organizations are already well acquainted with those ideas, as decision automation is now commonplace in high frequency of interaction contexts, such as consumer lending or insurance claim processing. That said, there is a yet another, often overlooked dimension of human-computer interaction (HCI), one where humans and computers are engaged in in a two-way interaction, termed here human-computer collaboration (HCC).
HCI sees the manner in which humans interact with machines as one-way phenomenon, primarily concerned with making computers and other devices as easily usable as possible. Clearly, that is a critical consideration when designing human-facing interfaces of computing systems, but it tacitly dismisses the idea of collaboration-like, two-way human-computer interactions. To put it in terms of organizational ecosystems, HCI frames computers as ‘suppliers’ of specific informational inputs, rather than ‘partners’ in the knowledge creation process. And while that framing is reasonable when computers serve as de facto high-end electronic calculators (which, to be sure, is still the core and tremendously valuable functionality of computing systems), it is too restrictive in situations in which advanced computing systems are capable of self-controlled intelligent behavior. Of course, as noted earlier, not being conscious, artificial computing systems are not (yet?) capable of truly independent ‘thought’, meaning that their informational output is ultimately constrained by the combination of their programming and data from which they learn. Still, those systems can process vast quantities of data tirelessly, accurately, and at a dizzying speed, all of which raises a tantalizing possibility: What if human informational workers and artificial computing systems could engage in teamwork-like collaborations? Could the current conception of ‘machine learning’ be expanded to also encompass adaptation of ANN (artificial neural networks) and other machine learning algorithms to the nuances of individual analysts’ areas of analytic focus? In other words, could machine learning entail not just algorithms extracting meaningful patterns out of data, but also learning about the informational needs of their human partners? All considered, what if the current conception of data scientists’ and business analysts’ skillset was to be expanded to not only include the traditional computing competencies, but also HCC-framed teamwork? Would that offer organizations the long sought-after ability to monetize their data assets?
The idea of monetizing organizational data is compelling to all, but to-date only a handful of organizations found effective ways of doing so, and most of those organizations, such as Google or Facebook, are themselves products of the Information Age. Most companies’ business models are not built atop of data flows, which means those organizations are not innately positioned to extract direct economic benefits out of their data holdings. For those organizations, the path to extracting economic value out of their data assets is through broad organizational embrace of the idea of data-enabled creativity. Operationally, that means going beyond the tried-and-true modes of data utilization, which nowadays encompass anything from outcome tracking and reporting mechanisms to predictive analytic capabilities. In the past, brilliant breakthroughs were often painted as random flashes of light in the dark of night, but going forward, breakthroughs could become far more predictable outcomes of systematic pursuit of discontinuous innovation, rooted in meaningful human-computer collaborations. There is nothing outlandish about such an assertion as it has long been known that innovation-supportive organizational practices and culture produce steady flows of cutting-edge innovation (perhaps best illustrated by the 3M Company’s 55,000+ products); HCC is merely the next step in that evolutionary development.
The New Frontier of Imagination
The preceding reasoning is rooted in an implicit assumption that once realized, a breakthrough idea will be recognize as such, but there are ample reasons to be skeptical. Why, for instance, are some works of art celebrated while others are looked at with disinterest, or barely noticed? Or why are the same works go unnoticed at first, only to become artistic superstars later? As is well-known, Vincent Van Gogh, the famed post-impressionist painter, created some of the most expensive paintings to have ever sold, he even has a major museum devoted entirely to his works (the Van Gogh Museum in Amsterdam), yet he did not enjoy any of that success during his lifetime, in spite of completing more than 800 paintings (more than 2,100 individual artworks in total) in a span of just a decade. What makes his Starry Night or his Irises so ‘creative’? Along similar lines of reasoning though in far less artistic sense, the category of modern sport utility vehicles, or SUVs, is commonly believed to have been sparked into existence by the introduction of 1990 Ford Explorer, a norm-setting design. Yet the 1984 Jeep Cherokee was a remarkably similar (both were 4-door, comparably sized and shaped vehicles with essentially the same or comparable attributes); why then does the Explorer get the creative glory?
Is some form of success required to earn the ‘creative’ badge? It is a bit like the familiar sounding ‘if a tree falls in a forest and no one is around to hear it, does it make a sound’ quandary. Questions like that do not lend themselves to straightforward, objective answers but there is value in considering them, and that value is the development of a meaningful perspective on which the idea of data-enabled creativity can be operationally developed.
Whether or not some form or success or wider acceptance are deemed necessary, creativity tends to be assessed through the prism of ‘stuff’, such as the above Van Gogh’s paintings or category-defining automobiles. In an organizational setting, that translates into output, be it in the form discrete products or services, or decision choices. That perspective, however, overlooks an increasingly important, latent aspect of creativity: situational sensemaking. The current highly volatile, hypercompetitive socio-politico-economic environment rewards organizations, seen here as human collectives working together in pursuit of shared commercial or societal goals, that exhibit superior abilities to problem-solve and to adapt to the ever-changing environmental realities. If volatility and change are ongoing and largely unpredictable, organizational ability to respond is ultimately contingent on individual organizational constituents’ situational sensemaking, or considered more broadly, creative problem-solving capabilities. In that particular context, creativity is construed to be the ability to identify emerging opportunities and threats, and adaptively respond to observed or perceived changes in one’s environment. While that framing certainly does not preclude innovative product or service ideation, it is nonetheless primarily focused on widening and deepening of individual-level cognitive thinking and emotive feeling capabilities, seen here as founding enablements of situational sensemaking. And lastly, it is a crucial element of the earlier discussed human-computer collaboration.
The above reasoning is rooted in the belief that, at its core, human creativity entails finding new and meaningful arrangements of already-known elements (rather than conjuring up otherworldly ideas); in a sense, it could be likened to finding new patterns in old data. An essential enabler of being able to do that is imagination, generally understood to be a process of forming new ideas or representations not present to the senses. While the difference between creativity and imagination is not always clear, the former is generally seen as a purposeful process, which means it tends to be confined by some pragmatic bounds, while the latter is usually neither purposeful nor confined by any type of bounds. Hence it is that unbounded nature of imagination that grants it the freedom to conjure up new concepts, perspectives or associations that are unlikely to be produced without it, and imagination is also the aspect of creative problem-solving that can be stimulated by means of computer-rendered simulations.
It is important to note that what is described here represents a fundamentally different conception of computer simulation-based learning than the one used to train pilots, nurses, or business managers, all of which are based on the idea of artificial rendering of accurately depicted physical reality, with the goal of emulating the conditions those professionals are expected to encounter on-the-job. The stated goal of data-enabled simulations discussed here is to do the opposite – to create scenarios, associations, or conditions that may not have yet been encountered, with the goal of stimulating imagination, and ultimately, creative problem-solving. Operationalizing of that idea, however, is contingent on closer alignment of human and machine learning processes, which in turn touches on the idea of ‘semantic triangle’, first suggested by Ogden and Richards nearly a century ago. It offers an abstract representation of how linguistic symbols are related to the objects they represent, and when framed in the contemporary context of human and machine learning, its core premise is graphically illustrated by Figure 1.
Figure 1
Semantic Triangle
As graphically depicted above, the primary sensemaking mechanism in human learning relies on associating symbols, such as words, with typically contextual meaning; on the other hand, the primary machine learning avenue relies on grouping of specific examples, or points of reference, into larger categories, which is seen as the ability to generalize. Stated differently, being self-aware, humans develop conscious understanding, while self-awareness-lacking artificial learning systems are not capable of developing of human-like understanding, framed here as cognitive sensemaking. Simply put, the association between, for instance, ‘task involvement’ and ‘result quality’ can be expected to evoke a particular meaning in a human, which can be seen as a product of an intuitively obvious, implied abstract relationship, easily grasped by most. To a machine, however, the same association is not ‘understood’ in terms of such elusive, transcendental qualities – instead, it merely belongs to a particular class of associations, which means that artificial intelligence manifests itself in the ability to create semantically and practically meaningful groupings, rather than human-like cognitive understanding. With that in mind, the idea of simulation-based learning is predicated upon the assumption that given the many potential classes of associations that machine learning systems can produce, at least some can be expected to fall completely outside of the evoked or consideration set of human meanings. In other words, by creating new but algorithmically plausible (i.e., consistent with embedded computational logic) classes, machine learning modality can suggest potential new ideas that reach beyond what a given human learner could be expected to conjure up on his or her own. In that sense, artificial systems-based simulations can help humans ‘imagine’ the otherwise unimaginable.
To summarize the preceding discussion, the expanded, data-enabled conception of creativity is captured in Figure 2.
Figure 2
Expanded Conception of Creativity
People, and thus organizations, have been learning from printed sources for the past 500 or so years, from mass audio sources such as radio for about 100 years, from audio-visual television for roughly 70 years, and computer-rendered online sources for about 25 years. Simulation-based learning is merely the next, logical step in the evolution of human cognitive development. One of the most celebrated scientists of the modern era, Albert Einstein, famously noted in a 1924 interview that ‘…imagination is more important than knowledge…knowledge is limited, whereas imagination embraces the entire world, stimulating progress, giving birth to evolution.’ Using technology to further stimulate the very human imagination that, unaided, has already produced so much, will undoubtedly propel mankind to creative heights that, at the moment, may seem, well, unimaginable…
Comments