COGNITIVE ARBITRAGE: THE OUTSOURCING OF INTELLIGENCE
Artificial intelligence (AI) is disrupting industry and potentially threatening to replace humans at work. In this article, we offer a strategy to ensure that executive decision-makers are given the tools to combine the best of human skills with AI, both preserving human dignity and enhancing organizational achievement. We propose a decision-making framework, the Arbitrage-Enhancement Decision Grid (AEDG), that enables organization leaders to determine the optimum human and intelligent machine collaboration to improve workforce performance. The framework recognizes the inevitable adoption of technology innovation, in conjunction with an organization’s need to balance human performance and competitive objectives. The authors then advance an actionable roadmap for developing human workforce and intelligent machine competencies and skills, the Human Resource-Artificial Intelligence Collaboration (HRAIC) framework that complements the decision-making outcomes of the AEDG.
The rise of the robots will boost productivity and economic growth. And it will lead to the creation of new jobs in yet-to-exist industries. But existing business models in many sectors will be seriously disrupted and millions of existing jobs will be lost.
You can pretty much buy any technology, but your ability to adapt to an even more digital future depends on developing the next generation of skills, closing the gap between talent supply and demand, and futureproofing your own and others’ potential.
Imagine, if you will, three contrasting scenarios. In the first scenario it is 2015. You ask Alexa for a suggestion on where to eat, Google Maps to direct the turns as you head to your destination, and your Apple Watch the number of steps needed after dinner to mitigate your caloric intake. You then ask Siri what might be fun and safe to do after dinner, just a short drive away.
The next scenario is the year 2025, and you use a single onboard Tesla application to map out your evening’s itinerary. You have more leisure time than usual this evening because ChatGPT is writing a draft of your training program, and a draft of your partner’s legal brief.
For the last scenario, a decade later, it is 2035. You can only imagine what you will see, feel, and sense with the aid of machines. The future of technology is perplexing, hopeful, and perhaps perilous in your business and personal life.
THE ARTIFICIAL INTELLIGENCE “BOILING FROG”
There is a palpable angst in the present that accompanies the rapid changes in our current cognitive revolution, led by the relentless march of artificial intelligence (AI) (Daron & Johnson, 2023; Leffer, 2023). Increasingly machine programming augments both our personal and industrial lives, sometimes in ways that simplify activity, and some with disruptive consequences to occupation, family, and geography (Acemoglu & Johnson, 2023; Poba-Nzaou, et al. 2021). Have we begun to relinquish aspects of our thinking and problem solving to the machines, literally and figuratively outsourcing our decision-making and portions of our intelligence?
To avoid future AI “boiling frog” scenarios, what if we could anticipate, that is, capture and articulate the changes, and begin to answer the following questions: “What are the changes?” “In what direction are the changes moving?” “What can we do, especially regarding the impacts on human labor?” “What are the talent requirements and what individual/organizational performance will be expected?”
In this article we use a variant of the term arbitrage. Arbitrage involves trading in two different markets to lock in profits. The authors propose a decision-making framework, the Arbitrage-Enhancement Decision Grid (AEDG), that enables organization leaders to determine optimum human and intelligent machine collaboration to improve performance. The framework recognizes the inevitable adoption of technology innovation, in conjunction with an organization’s need to balance human performance and competitive objectives. The authors then advance an actionable roadmap for developing human workforce and intelligent machine competencies and skills, the Human Resource-Artificial Intelligence Collaboration (HRAIC) framework, which complements the decision-making outcomes of the AEDG (Jain, et al. 2023; Yue & Li, 2021).
Foundational to executing the AEDG, the authors first introduce the term cognitive arbitrage to describe business preference to apply intelligent machine technology to automate human tasks. Following, the authors introduce the term cognitive enhancement to describe the benefits of innovative human—intelligent machine collaboration. In building an actionable roadmap of AI skills and competencies that define HRAIC, the authors refine the current definition of knowledge workers to differentiate between digital and analogue knowledge workers.
Our objectives are threefold:
-
Allay workforce fears related to AI while acknowledging the inevitable adoption of AI technology innovation, contrasting cognitive arbitrage with cognitive enhancement.
-
Provide a framework for organization leaders to anticipate change, and wisely plan for the AI future. We articulate decision-making rules for determining optimum human- machine innovation and collaboration as seen in the AEDG framework.
-
Describe talent competencies emphasizing human-machine collaboration. We seek a balance for optimum worker performance and engagement, and thus map AEDG analysis to the HRAIC talent profile framework. We express this by developing ideal talent personas.
CURRENT AND FUTURE AI WORKFORCE LANDSCAPE
Labor arbitrage involves disintegrating organizational barriers to place job positions where they are geographically less costly. This term cannot entirely capture the gut-wrenching complexity and rapid changes brought on by machine intelligence. Whereas mid to late 20th century computing advances augmented human tasks, early 21st century computing advances based on principles of cognition and AI are replacing human decision making and problem solving. Rapid technological advancement has enabled the pencil and paper evolution of neural networks, machine decision making and robotic ‘relationships’ to evolve from science fiction fantasy to Wall Street monetization at an accelerating pace (Hermann, 2023; Kelly, 2024).
To anticipate the current and future impact of rapidly developing machines and artificial intelligence, the authors introduce the notion of cognitive arbitrage. We define cognitive arbitrage as efforts to shift industrial business and human processes up higher levels of cognitive complexity to achieve lower costs and improve efficiencies (akin to labor arbitrage) (Boquen, 2023). Specifically, cognitive arbitrage escalates what were solely human cognitive tasks and skills to reside natively in a hardware/software interface, often transparent to the end user.
When economic advantages of labor arbitrage become either exhausted or unobtainable (e.g., outsourcing, offshoring, and/or nearshoring) organizations turn to labor enhancement. This form of labor employs innovative, collaborative worker solutions that retain workers while also improving business performance outcomes. Some examples are Six Sigma, Lean Manufacturing, Agile software development and Design Thinking (Roberts, 2019).
Labor and cognitive arbitrage differ from labor enhancement and what the authors term cognitive enhancement. Cognitive enhancement involves human-automation teaming (HAT) technologies that take advantage of humans’ cognitive strengths and the power of AI. The focus is collaboration in a way that uses the power of computing pattern recognition and speed with the experience of human creativity and experience. Some of the more common examples include robotic surgery and self-driving vehicles.
Our two strategic frameworks, the AEDG and HRAIC, pursue an ethical approach that we believe fosters innovation, job satisfaction, and a sense of participation. Using machines to replace drudgery makes sense, but social progress is not enhanced when machines are put in competition with people. We argue that firms are better served when they build roles where human workers and intelligent machines provide an optimum balance of performance, innovation, cost, and competitive advantage. The HRAIC identifies ideal ‘personas’ of talent competencies and skills that support HAT, helping to mitigate the emerging fear of machine-engineered redundancies.
Regarding the fear of shifting job roles and AI, Aslan & Acemoglu (2021) remark:
…innovation and participation cause employees to develop affection toward their jobs and organization, leading to the psychological experience of the organization’s ownership and, ultimately, job satisfaction (p. 10).
ECONOMIC AND ORGANIZATIONAL IMPACTS OF LABOR ARBITRAGE AND LABOR ENHANCEMENT
In the following section, the authors review the economic and organizational impacts of labor arbitrage and labor enhancement. The section following assesses the parallel impacts of artificial intelligence on arbitrage and enhancement, positioning the impacts of AI as cognitive arbitrage and cognitive enhancement. The authors then introduce the AEDG. The AEDG assists executives and management to select paths that promote the advantages of collaborative AI/human partnership. Finally, the authors introduce the HRAIC, and discuss the implications for talent, learning, and human performance.
Labor Arbitrage
Implementation of labor arbitrage involves several financial tools. Labor arbitrage is most closely associated with the practices of outsourcing and offshoring. Outsourcing involves contracting with a third party provider for services or products. Outsourcing may be contracted with a firm in a different country from the parent firm—termed “offshore outsourcing”—or can be with a firm in the same country but a lower cost region. This is termed ‘domestic outsourcing’. Parent firms may also retain direct management and legal control of their services and production (i.e., they do not contract with a third party), but move them to lower cost countries, termed simply “offshoring”; or move operations to a lower cost region within the same country, termed “nearshoring” (Willcocks, et al. 2015).
There are several economic drivers for outsourcing and offshoring, but reducing operational costs is typically a primary driver. Consultancy Allied Global (2023) states:
Outsourcing helps companies experience significant savings in cost reduction in the rates vendors offer outsourced employees compared to in-house employees. In 2018, 62% of companies reported 10% to 25% savings when they outsourced and the remaining 38% of companies reported savings as high as 40%.
Labor arbitrage has been an attractive means by which to control costs. However, the positive benefits for country, state, and regional geographies, not to mention displaced labor, remain elusive. Addressing the practice of labor arbitrage and offshoring Levy (2005) states:
…reducing wages through offshoring leads to wealth creation for shareholders but not necessarily for countries and employees, and that many displaced workers have difficulty ‘trading up’ to higher skilled jobs…it decouples the linkages between economic value creation and geographic location. The result is the creation of global commodity markets for particular skills and a shift in the balance of market power among firms, workers, and countries (p. 1).
Ndubisi & Nygaard (2018) also call into question the balance of economic benefits, social impact, and organizational performance from labor arbitrage. Should firms make business decisions based solely on profit motive, regardless of regional or individual consequences? A similar question applies to AI. What are the social and organizational impacts of ‘outsourcing’ to intelligent machines?
Labor Enhancement
Labor enhancement involves implementing a number of innovative labor practices and business process tools having measurable financial benefit to firms. In contrast with labor arbitrage, which seeks workers willing to perform similar work tasks at lower wages, labor enhancement seeks to change business practices and processes. In this case, workers are retained with positive impacts on morale. Enhancing business processes is rooted in a firm’s commitment to and execution of innovation. In an AI world, what does arbitrage vs. enhanced innovation, or enhancement for short, “look like”?
Roberts (2019) states:
Enhancement-oriented innovation often starts with the question of “How might we do X better?” It is not about questioning what is being done, but rather how it is done and whether it can be done differently, and hopefully better. Common methods or practices that underpin enhancement-oriented innovation are generally structured learning processes to help consolidate insights and build upon them. Example practices include lean, business process management, quality control, and behavioral insights.
According to the European Central Bank (2017), the core economic driver of innovation is financial growth through increased productivity, involving the same worker(s) producing higher output, that is:
… the major benefit of innovation is its contribution to economic growth. Simply put, innovation can lead to higher productivity, meaning that the same input generates a greater output. As productivity rises, more goods and services are produced.
Two specific examples of labor enhancement, TQM and Agile, illustrate the economic and labor impact of each innovation.
Total Quality Management (TQM)
TQM is an overarching term that refers to multiple quality tools and methods developed to improve product and service process output. Some examples include Deming’s (1986) method of statistical analysis, Crosby’s (1979) promotion of “zero defects” to eliminate rework; and Ishikawa’s (1976) “fishbone” causality analysis. Today, these and other methods and tools are nested within the global quality standard referred to as International Organization for Standards ISO 9001 (2015).
In a recent meta-study of 172 articles addressing quality management system implementation, García-Fernández, et al. (2022) found:
… quality management practices have positive effects on operational and financial performance. Most of the papers show a positive relationship between quality management and financial performance (22), and between quality management and operational performance (13). Others show a relationship between quality and both types of performance: operational and financial (4) (p. 12).
Research of quality management systems reveals positive impacts to worker morale in multiple areas. Karia and Asaari (2006) note, “Managers should be aware that TQM practices have a positive effect on employees' work‐related attitudes…with significant positive effect on job involvement, job satisfaction, and organizational commitment” (p. 1). Pratiwi, et al. (2019) observed “Employee morale and achievement of company performance will improve aligned with the implementation of quality management supported by the company”.
Agile Software Development
Agile is an example of technology driven labor enhanced innovation. Like TQM, it has become an inclusive term that includes a variety of methods spawning new job roles, for example Kanban, Scrum, DevOps, and others. At its core, Agile is an iterative process for developing and implementing software applications (Dingsøyr, et al. 2012).
With respect to economic drivers of this example of labor enhancement, in a meta-study of over 300 articles on Agile methodology (Rico, 2009) found evidence of multiple advantages.
On average, studies of Agile Methods reported 29% better cost, 91% better schedule, 97% better productivity, 50% better quality, 400% better satisfaction, and 470% better ROI.
While Agile economic incentives continue to drive adoption, Agile implementation benefits workers. Rigby, et el. (2018) note: “When implemented correctly, agile innovation teams almost always result in higher team productivity and morale, faster time to market, better quality, and lower risk than traditional approaches can achieve”.
As specific examples of labor enhanced innovation, practices like TQM and Agile development—unlike labor arbitrage influenced practices—foster improved morale, job satisfaction, and worker performance without massive job disruption and/or elimination.
COGNITIVE ARBITRAGE AND ENHANCEMENT: THE EVOLUTION AND IMPACT OF INTELLIGENT MACHINE LABOR
Any current and future theory of labor arbitrage and enhancement begs the inclusion of cognition in describing labor practices. This is due to the increasing need for knowledge workers, or those individuals with the depth of cognitive and analytical skills necessary to perform 21st century work tasks. The term knowledge worker, in its most general sense, means a worker who adds value to the product or service (e.g., data, information, interpretation, cross-referencing, or other forms of conceptual processing) (Koraeus & Stern, 2013). Stotler (2019), notes:
… nearly half of U.S. employees alone may now be defined as knowledge workers. These workers must rely on cognitive skills and analytical and critical thinking to get their jobs done.
The following section addresses concerns about the impacts of AI on labor arbitrage and labor enhancement to assess knowledge workers’ future needs in industry. We describe these impacts as a decision between either cognitive arbitrage or cognitive enhancement.
Cognitive Arbitrage
Recall that cognitive arbitrage is a shift of human labor to technology, whereas labor arbitrage shifts labor to a new geography. Some examples cognitive arbitrage would include ATMs, digital assistants, robotics, natural language processing, chatbots and self-driving vehicles.
In labor arbitrage, jobs are outsourced and/or offshored to other (human) laborers. In the case of using AI for cognitive arbitrage, many now human tasks are on a trajectory to eliminate humans’ involvement and reside in software and hardware. A World Economic Forum (WEF) Future of Jobs Report (2020) estimates that 85 million jobs will be displaced due to AI by 2025. The number of jobs lost to cognitive arbitrage potentially outstrips the number of jobs lost to labor arbitrage. Though the WEF estimates that in the same period 97 million new jobs may be created, this represents a magnitude of labor disruption and change management challenges.
Addressing the rise of AI-enabled robotics, Oxford Economics (2019) reports:
The rise of the robots will boost productivity and economic growth. And it will lead to the creation of new jobs in yet-to-exist industries. But existing business models in many sectors will be seriously disrupted and millions of existing jobs will be lost. We estimate up to 20 million manufacturing jobs are set to be lost to robots by 2030. The effects of these job losses will vary greatly across countries and regions, with a disproportionate toll on lower-skilled workers and on poorer local economies. In lower-skilled regions, we find that robots lead to almost twice as many manufacturing job losses. In many places, the impact will aggravate social and economic stress in times when political polarization is a worrying trend.
The costs of worker displacement are jarring and require organizations to develop competency-building training programs for workers to overcome the challenges we can anticipate. In a study (Green, 2020) of AI and ethics, The Markkula Center for Applied Ethics determined that AI-induced unemployment will result in a number of issues. These include a loss of meaning; socio-economic inequality disruption; fears of AI consciousness; robot rights and personhood; and deskilling and debility (that is, loss of human competence).
Like labor arbitrage, cognitive arbitrage may appear on the surface to be an attractive option for executives to control cost. It lowers costs via employing fewer human workers as opposed to labor arbitrage which employs less costly human workers. In many cases machines can reduce the error and increase productivity when compared to human-only labor. However, the cost of support (e.g., upskilling, universal basic income, etc.) for millions of those displaced calls into question the long-term benefit of cognitive arbitrage (Oxford Economics, 2019).
The authors conclude that the practice of cognitive arbitrage, akin to labor arbitrage, brings into question the impact of purely economic driven benefits. Firms are cooperatives of people who serve a purpose in society. Like machines, firms must be efficient, effective, and profitable, but firms and their people are more than “things”.
In the following section, the authors propose that cognitive enhancement, like labor enhancement, provides organizations a balance of innovation driven economics and performance improvement with worker purpose and value.
Cognitive Enhancement
Cognitive enhancement, like labor enhancement, involves introducing innovative technology that works alongside the human (e.g., AI, HAT) to increase engagement and improve performance and profit. By contrast to cognitive arbitrage, where human tasks are delegated entirely to machines, cognitive enhancement focuses on collaboration. AI-augmented labor changes business processes in a way that retains workers and offloads the repetitiveness of some work elements to the machines (Tariq, et al. 2021). The objective of cognitive enhancement is to upskill human labor with the assistance of AI. This improves job participation, morale and performance. Li, et al. (2022) note:
…if AI is allowed to undertake analytical computational tasks in cooperation with humans, its performance is likely to be enhanced substantially … (p. 1).
Mallick, et al. (2023) go on to state:
… findings show that integrating emotions within AI teammates has a positive influence on human perceptions and behavior in a task.
What is cognitive enhancement in practice? We describe two examples of AI-augmented human labor below, medical diagnosis and financial services.
Medical Diagnosis
AI-augmented medical diagnosis uses AI algorithms in conjunction with a variety of medical (e.g., imaging) diagnostic tools such as deep (learning) neural networks (DNN) (Kim, et al. 2019). Human interpretation of imaging technologies (e.g., X-rays, ultrasound, nuclear medicine, positron-emission tomography, magnetic resonance imaging, mammography, and computed tomography) is now aided by AI. The diagnosis of many maladies—Alzheimer’s, cancer, diabetes, chronic heart disease, tuberculosis, stroke and cerebrovascular, hypertension, skin, and liver disease (Kumar, et al. 2023)—can be improved by AI. AI adoption in the general healthcare marketplace is projected to grow from US$2 billion in 2018 to US$36 billion by 2025 (Hazarika, 2020). Collier & Fu (2020) project that AI could account for up to $3 billion in annual savings for the US health care economy by 2026 through improved imaging diagnosis. The use of an AI/cognitive enhancement approach to pathology improves patient outcomes but it also addresses a global shortage of pathology specialists. Kumar, et al. (2023) state:
… AI can result in improved patient outcomes and increased productivity and efficiency in care delivery. It can also enhance healthcare practitioners’ daily lives by spending more time caring for patients, therefore increasing staff morale and retention (pp. 8475–8478).
Financial Services
AI-augmented financial services encompass a wide variety of financial added-value processes. In banking, ATMs replaced much of tellers’ workload. It was feared this would eliminate banking position, but the opposite has been the result as banks added new value-added personal services that require more skills. As is the case for medical diagnosis, financial services can apply AI to enhance the financial experience in ways that combine technology and experienced bankers to improve customer satisfaction (Fares, et al. 2023). Specific applications are credit and risk underwriting, loan approval, fraud detection, wealth management, and blockchain (Fernandez, 2019; Verbeek & Lundqvist, 2021).
There are compelling financial incentives for firms to adopt AI-augmented financial services. By 2035 the banking sector could see an overall productivity increase of 4.3% (Bredt, 2019). In a review and study of the banking sector, Fares et al. (2023) note that “banks are expected to save $447 billion by 2023, by employing AI applications.” Successful AI augmentation of the financial services industry encourages the blend of AI with what Mehrotra (2019) terms the human/personal touch. He states:
… emerging capabilities of AI are combined, re-constituted and re-formulated in unexpected ways and are throwing up new opportunities and new challenges but at the same time posing new threats also…a major question to be inquired into is its sustainability as it tends to replace humans and the related personal touch which most often is the essence of financial services industry thriving on the art of customization and customer delight.
As examples of cognitive enhancement, healthcare and financial services can address the issues of consumer care, employment risk, and talent creation. The purposeful upskilling of repetitive roles can mitigate job reduction risks and develop the human talent necessary to partner with AI augmented finance and medical care for improved performance.
THE AEDG
This section describes and elaborates the previously mentioned AEDG framework (Mea & Eicher, 2021), a decision-making tool to enable organization leaders to determine optimum organizational planning for future human and intelligent machine collaboration. We use it also to refute the zero-sum game viewpoint articulated by De Cremer and Kasparov (2021) that there can be only one winner in the future: either humans or machines, but not both. The zero sum POV assumes that intelligent machines can model human capabilities, and that automation of those same capabilities will be superior to human only capabilities.
We propose assessing the relative impact of arbitrage and enhancement along two dimensions: cost and automation. Cost captures economic decision considerations, and automation captures the evolving blend of human skills and machine capabilities. We believe that modeling these two factors helps strategic planning in a way that anticipates business model disruption. It helps point out where arbitrage may make sense and where enhancement will lead to the greatest benefits. HAT shows great promise when it comes to profit, worker engagement, customer, and staff satisfaction.
As depicted in Figure 1, cost is represented on the X-axis and automation on the Y-axis. Each of four quadrants is associated with our prior discussion of labor arbitrage, labor enhancement, cognitive arbitrage, or cognitive enhancement. In order,
Citation: Performance Improvement Journal 63, 2; 10.56811/PFI-24-0015
-
Quadrant I requires higher investments with interactive machine capabilities, representing Cognitive Enhancement. In this case (e.g., medical diagnostics and financial services), technology supports the human in control, providing optimum innovation through human/AI partnership, upskilling, and collaboration.
-
Quadrant II represents Cognitive Arbitrage with low costs and high automation, in which machines take over many of the day-to-day human duties. In effect, human labor and thinking are ‘outsourced’ to machines that require no supervision and limited maintenance.
-
Quadrant III, with low automation and low costs, represents Labor Arbitrage, driven by the outsourcing and/or offshoring/nearshoring of human labor to low-cost geographies.
-
Quadrant IV, with low automation and higher investment in people, represents Labor Enhancement, where human labor is retained and upskilled using organizational development tools to improved business processes, and support greater personal autonomy and teaming.
Figure 2 below depicts the AEDG framework to include industry examples illustrative of where various products and services may reside. Developing a similar grid like this is helpful in HR strategy as it makes where roles might reside and signal where improvements can be made. Note that the authors’ examples are directional and evolving for the purposes of debate and discussion, potentially encouraging informed decision-making. Each of the products and services listed is ever evolving in a dynamic AI marketplace.
Citation: Performance Improvement Journal 63, 2; 10.56811/PFI-24-0015
Enhancement Quadrants I and IV on the right present a contrast to Arbitrage Quadrants II and III on the left. The benefits of enhanced innovation, both cognitive and labor, suggest the pursuit of collaborative enhanced innovation with intelligent machines. This addresses the problem with outsourcing human decision making solely to intelligent machines and can provide measurable worker and organization benefits.
The authors intentionally emphasize Quadrant I products and services as prototypical examples of human-enhancing HAT collaboration. Examples of the benefits of human worker cooperation with intelligent machines are discussed by Ramchurn, et al. (2021); Li, et al. (2022); Jain, et al. (2023); and Tariq, et al. (2021) and are the foundation for creating a decision-making practice that informs the talent requirements of an AI workforce.
TALENT IN THE AGE OF INTELLIGENT MACHINES
In order to plan for the inevitable realignment of future human resources in age of AI, we need to ask: What are the requisite skills and competencies necessary to optimize human labor and intelligent machine performance? The authors advance an actionable roadmap below for developing human workforce and intelligent machine collaboration set of competencies and skills that we call the HRAIC framework. It is a complement to the strategy decision-making process in the AEDG. The authors’ intent is to provide HR leaders with a guide to vet talent considerations that support leaders in the age of intelligent machines.
When executive management chooses to enhance, this pushes the levers of innovation to increase the value of its people. Weekes and Eskridge (2022) state, “Over the past decade, the demand for high-performing knowledge workers has grown at an unprecedented rate and shows no signs of slowing.” Anticipating knowledge worker competencies requires the sort of skill definition that organizations sometimes ignore. A future roadmap is needed.
Toward building an actionable roadmap in the HRAIC skills we differentiate between digital and analogue knowledge workers. Moskaliuk, et al. (2017) articulated these as “Typical (knowledge worker) tasks (involve) … for example, acquiring, searching, analyzing, and storing knowledge, as well as organizing, planning, and deciding.”
-
Digital knowledge worker’s tasks involve a focus on search, acquisition, analysis, and storing data, with a bias toward technical competencies and skills in support of AI related cognitive practices.
-
Analogue knowledge worker’s tasks involve organization planning and decision making, with a bias to interpersonal competencies and skills in support of non-intelligent machine labor practices.
As Roberts (2019) earlier noted, successful enhancement demands innovative or breakthrough thinking to create novel outcomes. Capponi, et al. (2022), in a study of breakthrough innovations states:
Breakthrough innovations significantly depart from common practices and can potentially reshape existing markets, create new markets, and prompt the emergence of new technological trajectories (p. 1).
Labor and cognitive enhancement align along a Knowledge Worker Maturity Scale as depicted in Figure 3 below. It depicts the relationship between digital and analogue workers. Defining digital and analogue knowledge worker skills helps bolster opportunity for enhancing future HR planning and breakthrough opportunities. Arbitrage actions only commoditize firm processes and lower immediate costs, but perhaps to the detriment of future success. Figure 3 highlights example results of decision-making using the AEDG framework. This stresses the difference in the potential benefits of breakthrough skills and implies a maturer HR model that scales toward human capabilities.
Citation: Performance Improvement Journal 63, 2; 10.56811/PFI-24-0015
TALENT, AI, AND PERFORMANCE
In our assessment, human talent in collaboration with technology, will drive success in the AI future. In support of this position, a recent survey by David (2023) states,
AI is not IT. The skills and knowledge that allow the organization’s IT workforce to run and maintain more traditional technology platforms are not the same as those required for AI programs…Neither technology nor talent acquisition can be decided in a vacuum. These two sides of the same coin affect and influence one another, and neither can function without the other.
In the recent past, novel digital and analogue worker developments created new and sometimes controversial job roles. Position examples for digital workers would include Data Steward and Solution Architect. For analogue workers this would include Scrum Master and Six Sigma Black Belt. We propose a set of skills and competencies, that when combined, suggest an ideal AI worker persona. Compelling justification for identifying an AI worker persona can be found in research examining the impact of AI talent on organizational economic performance. For instance, in a study on ROI, talent and AI, Rock (2022) states:
… evidence that the expected future proliferation of AI talent causes previously sidelined AI projects to become profitable and existing AI projects to become even more profitable (“price effects”). The principal finding is that these price effects at the outset of a skill proliferation event are a likely mechanism than contemporaneous productivity increases or overall firm-level worker exposure to AI. These effects increase the value of installed capital that is complementary to AI talent (p. 5).
This study and others (e.g., Babina, et al. 2024) affirm that applying the optimum talent profile and working in collaboration with machines, improves organizational performance. Based on skill and valuation, Rock (2022) further states:
…the promise of this new AI is that it will lead to (often firm-specific) business process innovation, job redesign, automation, and new engineering advances across many domains in the economy (Furman & Seamans 2019; Brynjolfsson, et al. 2018; Felten, et al. 2018; Webb, 2019; Ransbotham et al. 2019). Even the relatively early bespoke applications of deep learning could feasibly cause large shifts in labor demand and economic value creation processes (Brynjolfsson, et al. 2018) (p. 8).
Frankiewicz and Chamorro-Premuzic (2020) go on to emphasize both critical talent and collaboration with innovative technology ensures future organizational performance as follows:
You can pretty much buy any technology, but your ability to adapt to an even more digital future depends on developing the next generation of skills, closing the gap between talent supply and demand, and futureproofing your own and others’ potential.
Towards an AI Worker “Persona”
We offer the following summaries of what an HRAIC model would ‘look like’ for prototypes of digital and analogue AI talent skills and competencies. Below is the author’s articulation of the minimal AI human capital skills and competencies required for successful human—intelligent machine collaboration. These categories of digital and analogue knowledge worker skills and competencies form the foundation of the HRAIC framework, complementing the decision-making outcomes of the AEDG. The categories are aspirational and will require further detailing for a specific organization’s requirements. Table 1 below provides a brief description of each Talent Profile.
We assert that ethically oriented and operationally competent managers need to develop future AI scenarios that incorporate a pivot in the HR understanding of talent. The interface between the human intelligence and machine operations now requires a new perspective, one that considers digital and analogue features to create novel roles. The following are a few composite persona skills:
-
Centripetal (orbiting toward the center) Thinking: Offering multiple, combinatory solutions.
-
Consummate Emotional Intelligence: Ability to “backchannel” irregular communications; empathize with both person and appreciate machine intelligence to weave comprehensible and ethical stories.
-
AI Application Fluency: Capacity to understand pairing AI design requirements with applications necessary to operationalize design.
Collaboration requires connection. The skills and competencies necessary to successfully collaborate with AI, that is, improve human and machine performance, might also include the overarching persona of a “cognitive helmsman,” that is, a navigator of the unexpected. Frankiewicz and Chamorro-Premuzic (2020) state:
It’s really quite simple: the most brilliant innovation is irrelevant if we are not skilled enough to use it; and even the most impressive human minds will become less useful if they don’t team up with technology.
Figure 4 below represents an integration of the AEDG and HRAIC frameworks into a talent matrix. Related matrices include taxonomies developed by Simmler and Frischknecht (2021); Liu, et al. (2023); and Mason, et al. (2023).
Citation: Performance Improvement Journal 63, 2; 10.56811/PFI-24-0015
The authors believe something like the HRAIC talent persona is necessary for organization talent leaders to manage the potential disruptions—economic, social, and ethical—of the constant pull toward short-sighted cost cutting of cognitive arbitrage. The skills and competencies necessary to successfully collaborate with AI may require a unique balance of upskilling, that is, learning new job skills in preparation for new roles; and reskilling, that is, improving current skills within the context of one’s current job role (Chakma & Chaijinda, 2020). This approach is well-supported (Kelly, 2024; Lang & Triantoro, 2022; Li, et al., 2022; Morandini et al., 2023; Sanders & Wood, 2023; Sawant, et al., 2022; Tamayo et al., 2023). Table 2 below summarizes how a push in the direction of each quadrant of the AEDG framework would potentially impact jobs and HRAIC talent requirements, with a focus on both upskilling and reskilling.
A key assumption of our frameworks is that wise executives and managers will pursue enhancement strategies, especially toward the goal of HAT (human-automation teaming) that takes advantage of the best combination of human creativity and AI algorithms. While there is risk and cost associated with moving in this direction, firms that pursue this direction can, over time, recoup research and development costs. In parallel, HR leaders need to support their organizations by developing job roles that keep them competitive for the future. By contrast, arbitrage brings short-term cost reduction that may not sustain a business entity in the longer term.
SUMMARY & CONCLUSION: FROM THE “GIG” TO THE “GIGABYTE” ECONOMY—WHO ARE WE NOW?
In summary, we have developed frameworks that support a process whereby executives, managers, and HR can make clearer decisions about the future in which AI technology can either replace or enhance workers. We:
-
Defined the terms cognitive arbitrage and cognitive enhancement to explore ways in which repetitive tasks either replace human with AI capabilities, or on the other hand, augment human with AI capabilities while enhancing productivity.
-
Articulated decision-making rules for determining optimum human-machine innovation, collaboration, and performance. The AEDG framework provides an approach where organization leaders can anticipate change and wisely plan for the AI future.
-
Mapped AEDG framework analysis to HRAIC skill profiles, describing talent competencies and requirements that emphasize human- machine collaboration in a way that balances performance and engagement.
Resistance to AI development often resides in the question of the autonomy of human decision-making and production being ‘outsourced’ to a machine. Resistance is compounded by the speed in which these changes occur, stripping individuals of autonomy and forcing them to recalibrate mentally and emotionally. In the future, organizations need to pursue pragmatic and ethical strategies that take advantage of the best that people and machines have to offer in combination with each other.
Cognitive and AI development is a rapidly moving target and may be shifting faster than humans can retrain, relocate families and lives, and balance local and global economies. We hope our approach can be one of many frameworks that mitigate cognitive arbitrage disruption. We believe our emphasis on cognitive enhancement—as opposed to cognitive arbitrage—provides a better long-term human/AI solution in the intelligent machine age.
AEDG Quadrants
AEDG Directional Examples
Knowledge Worker Maturity Scale
AEDG and HRAIC Talent Persona: “Cognitive Helmsman”
Contributor Notes
JAMES P. EICHER (MA) is the Founder and Managing Principle of Cognitive Management, and the author of many organizational change articles, management assessments and book chapters, as well the texts Making the Message Clear: Communicating for Business and Ecology of Truth: How Versions of the Truth Influence Behavior. He has held leadership positions at KPMG, Booz Allen Hamilton, Symantec, and IBM. Additionally, while an undergraduate student at the University of California, Santa Cruz (UCSC) he was both Dr. John Grinder’s and the late Gregory Bateson’s teaching assistant. Email: ecologyoftruth@gmail.com
WILLIAM J. MEA (PhD) is the Senior Director of Research at PsycHealth, LLC, which focuses on healthcare solutions and AI research. At Georgetown University, he teaches policy courses that draw from business best practices such as innovation, change management, program evaluation, and leadership. He served as a clinical psychologist in the U.S. Navy closing out his career in Fallujah. He was a manager at KPMG and Deputy Assistant Secretary at the U.S. Department of Labor before serving at the White House Office of Management & Budget. His research interests cover diverse topics in management, human behavior, AI, measurement, and ethics. Email: wm547@georgetown.edu
