AI, digital transformation, innovation, amplification: these are lofty, high-sounding buzzwords gaining more and more popularity. From startups turning to E-commerce, to stores launching apps, to boardroom discussions, to company town hall announcements, we hear these grand announcements and really expect to see something mind-blowing. Then, we wait and wait, bracing for impact, ready to see profits come pouring in－but then wait some more. Then, one begins to entertain, as though existential questions, whether the company was ready for it, when will it be ready? Is AI right for the business? Should have the company allowed more time for AI to develop further?
What is AI?
In the 1950s, Minsky and McCarthy coined the term “artificial intelligence.” Broadly, they defined it as “any task performed by a machine that would have previously been considered to require human intelligence.” Futuristic sci-fi movies imagine AI in sleek and crisp environments either in a utopia or in a dystopia. In effect, sometimes, these depictions elicit ambivalence and apprehension across a wide spectrum from the general public to theorists and academics, to wary practitioners and investors.
A modern view of AI takes on a tame, specific definition. For François Chollet, software engineer and AI researcher at Google, it is the intelligence tied to a system’s ability to adapt and improvise in a new environment, to generate its knowledge and apply it to unfamiliar scenarios. It is not a skill itself, not what you can do, but it is the efficiency with which you acquire new skills at tasks you didn’t previously prepare for.
Now, both these broad and specific definitions lead to two types of AI, one that is general, and one that is of the narrow type. General AI is the adaptable and flexible form of intelligence capable of learning how to carry out vastly different tasks. Basically, general AI is what we see in sci-fi movies.
In the 1980s, though, Moravec, Brooks, and Minsky articulated what is now in AI called the Moravec paradox. Hans Moravec (1988) observed that it is relatively easier to make computers perform well on intelligence tests but difficult or impossible to give them the perception and mobility skills of a one-year-old-child.
There are now, however, intelligent systems, such as those in computers that have become quite excellent at taking over specific functions from humans. While it is still a long shot before we can see artificial general intelligence being able to actually replicate the highly nuanced human intelligence, what we currently see in development is a more workable type: narrow AI. “Instead of building machines that are capable of doing everything, technologists focus on developing several narrow AI applications where machines outperform humans” (Marketing 5.0, 2021). In May 2020, François Chollet, @fchollet, tweeted that the thing about AI is you get what you optimize for. “If you optimize for a specific skill, like chess or StarCraft, your final system will possess this skill and nothing else. It won’t generalize to any other task.”
AI and Intelligence Amplification
While we need AI to make our tasks easier, it’s clear that AI also needs human guidance to properly learn and develop. It is in this sense that a collaboration between man and machine becomes beneficial for both towards learning what else can be achieved in future collaborations for each of their vision of intelligence amplification. “Knowing precisely what and how to teach computers will enable human coaches to realize their full potential,” leading “to a technology movement known as intelligence amplification (IA).” Kotler et al. (2021) differentiate AI and IA. “As opposed to artificial intelligence (AI), which aims to replicate human intelligence, IA seeks to augment human intelligence with technology. In IA, humans remain the ones making decisions, albeit supported by robust computational analysis.”
Man and machine: collaborating with AI
Ross and Taylor, in a November 2021 Harvard Business Review article, recommended four main AI management models. According to them, it is important to recognize this as a spectrum. Thus, their developed AI management models vary based on the level and nature of the human intervention. These are HITL, HITLFE, HOTL, and HOOTL.
Human in the loop (HITL). This is where a human is assisted by a machine. The human does the decision making and the machine only provides “decision support or partial automation of some decision, or parts of decisions. This is often referred to as intelligence amplification (IA).
Moreover, we can continue to wish that AI can have the ability to automate all decisions. Well, although it can identify patterns, provide simulations, and predict models from unstructured data sets, AI still needs human judgment, especially when it comes to qualifying deviations. These are the concerns of the second and third recommended models. Human in the loop for exceptions (HITLFE) is wherein most decisions are automated and the human only handles the exceptions. On the other hand, human on the loop (HOTL) is when the machine is assisted by a human. Here, “the machine makes the micro-decisions, but the human reviews the decision outcomes and can adjust rules and parameters for future decisions.”
Their fourth recommendation is the human out of the loop (HOOTL) model, wherein the machine is monitored by the human. The goal here is to make the machine decide every time. “The human intervenes only by wetting new constraints and objectives.” Here, improvements and adjustments based on human feedback are implemented into the automated closed loop.
Takeaway: intelligence amplification for digital transformation
Now, it is up to you to decide which among the above recommendations would suit your business. Caution, as in the aforementioned ambivalence and apprehension, is still important in guiding AI, even when it is for man’s intelligence amplification. As in Moravec’s paradox above, there are skills we just can’t transfer to machines. There is that thing called wisdom, for example: that which is sharpened from a wealth of practical－not theoretical－experiences.
AI, at least for now, is not a plug-and-play or a one-size-fits-all system. AI, digital transformation, innovation, amplification－yes, they are still lofty, high-sounding buzzwords. But, AI needs time to learn and develop. If you want, for example, to innovate, to adopt digital transformation through AI, then you simply have to start now (if you haven’t yet) and just allow it to evolve under an AI management model. What you can most certainly gain out of it is a developed ability to deliver a tech-empowered human interaction. When omnichannel presence, seamless activities, prompt responses, and quick improvements and upgrades are currently all in demand, a collaborative approach to working with AI for intelligence amplification is what could get you through to turning your business future-ready.
Kotler, P., Kartajaya, H., and Setiawan, I. (2021). Marketing 5.0: Technology for Humanity. John Wiley & Sons, Inc.
Moravec, H. (1988). Mind Children. Harvard University Press.
StratAccess Inc., established in 2012, commits itself to find its clients the right BPO for successful business solutions. The company focuses on transforming the landscape of business partnerships, especially in the need for digital transformation. StratAccess consultants stand ready to help clients take a hard look at their business objectives, organization infrastructure, and operational practices.
We at StratAccess strive to build long-term relationships that extend beyond the typical vendor-client transactions. Our primary focus is to successfully promote and serve each client’s products or services as though they are our own. Combined with the skill and knowledge of the outsourcing industry, our company has positioned itself as a leader in delivering its clients access to qualified quality and cost-effective BPO referrals.
Ivan Deligero is a contributing author at StratAccess. He likes deep dives into the bottom of things and sharing discoveries and strategies towards desired goals. His years of exposure in different industries have led to a deeper insight into organizational structures and operations, as well as the importance of process improvement. In his free time, he also reads and writes about some recent thoughts in philosophy.