> Can AI Enhance Students' Learning and Retention?

Walter Rodriguez, Ph.D., P.E.

Certainly! By combining cognitive science and artificial intelligence (AI), instructors can help their STEM students to learn & retain more of what they have learned, as we mentioned in our article "Can Machine Learning Enhance Human Learning in Times of Disruption?" (Rodriguez W., Angle, P., Snyder, M. 2021).

The AI/Natural Language Processing (NLP) learning & memory-retention authoring platform shown in the link below allows instructors to scan any text (or video transcription) and then the system suggests key concepts or important content that you might want to reinforce in your students and even assist with follow-up material creation. Therefore, it helps students retain more for a longer time: https://lnkd.in/g3_qdSS

…….

Can Machine Learning Enhance Human Learning in Times of Disruption? (Synopsis)

Full formatted published paper is available at https://cgscholar.com/bookstore/works/can-machine-learning-enhance-human-learning-in-times-of-disruption?category_id=cgrn&path=cgrn%2F216%2F217

Walter Rodriguez, Coursewell.com, USA

Patricia Angle, University of Oklahoma, USA

Michele Snyder, Florida Gulf Coast University, USA

Summary: In times of disruption, artificial intelligence, particularly the ubiquitous machine learning, is being rapidly integrated into online working environments in academia and industry. Primarily, these transformative technologies can help automate and perform continuous routine tasks using speed, efficiency, and effectiveness to gather and analyze massive amounts of data and information. This research illustrates ways in which machine learning can benefit learners and organizations---by making learners and workers work smarter and, more importantly, improve their ubiquitous learning environment anytime, anyplace.

Humans can learn while working and studying by integrating both reinforcement learning and supervised learning paradigms, and by augmenting, supplementing, or complementing machine learning with workers' humanness, skills, creativity, emotions, passions, and tacit knowledge. Furthermore, machine learning can track the humans learning process by learning from its experience--observing and analyzing the knowledge acquisition process, progress, and difficulties. And machine learning algorithms and agents can suggest new learning paths and approaches.

In addition to exploring if and how machine learning can enhance human learning, this paper investigates how it can assist learning without impinging on natural human and social learning development. Rather than just substituting or supplementing the human recall, machine learning can integrate and enhance work performance by minimizing---instead of supplanting---tedious tasks with automation and artificial intelligence.

Keywords: Machine Learning, Human Learning, Artificial Intelligence, Transformative Learning, Ubiquitous Learning, Reinforcement Learning, Supervised Learning, Information, and Communication Technologies

Introduction: In 2020, the COVID-19 virus was associated with a sharp increase in unemployment claims, as social (physical) distancing measures were implemented across the globe. While many employees were furloughed, some employers adapted by allowing employees to work from home using information and communication technologies. Many students and faculty were also impacted, and they too had to study and teach remotely. This paper shows how humans can learn and create value in work environments using artificial intelligence (AI) and, more specifically, machine learning (ML).

This research paper asks: Can ML enhance human learning (HL)? And explores how ML is already enhancing on-the-job training, leading to work performance improvements such as process streamlining, collaboration, and co-invention. Further, the study demonstrates how humans who embrace machine learning in the workplace and how employees who continue to learn and create value to the business environment (while being more productive, efficient, and useful) increase the chances of staying employed in times of disruptive change (i.e., automation, pandemics, hurricanes, earthquakes, environmental degradation, business, and social transformation, among others.)

Researchers envision humans and machines collaborating and being co-dependent, to bridge the gap between human learning and machine learning. To conceptualize the relationship between time and embedded constructs for humans and computers, imagine a relatively steep slope of the machine learning curve, compared with the human learning curve which illustrates the shorter time frame between the introduction of a dataset and AI learning. And, in the absence of a dataset, ML tools will learn from experience and associative rules and techniques via ML's reinforcement, supervised and unsupervised learning.

Over time, researchers theorize that the gap between associations, patterns and rules, and ML algorithms and agents will be formed. Also, associations formed by humans will grow increasingly faster. Of course, the certainty of it happening is subjective until rigorous empirical studies are conducted. Presently, humans and machines work together in what is called collaborative intelligence, but there is not yet evidence that human potential is increased by such collaboration (Daugherty and Wilson, 2018). If ML is applied in a collaborative manner with the human learner, rather than directly to perform tasks, is it possible to raise the speed of human learning and thereby decrease the gap between the two curves? Humans’ retention of information will lag behind machines since ML can be powered with cloud services and supercomputers using vast stores of knowledge that are difficult for humans to retain. Human recall can degrade due to forgetfulness even with some reinforcement assisted by AI-powered machines.

Historical Context: Machine learning is an undeniable part of human life for the past decade. The notion that automation (i.e., AI-powered robots) could be built and used to assist humans has been an idea brewing in the minds of science fiction authors since the early 1940s. Issac Asimov is the first to have published a story in which its plot revolves around robots called Runaround. Around the same time, during the Second World War, British mathematicians began developing code-breaking machines to decipher the enigma that the German soldiers used (Haenlein 2019). Almost one hundred years later, we now use computers in every specialty in the business world. But, is artificial intelligence, indeed increasing human productivity levels, cost-effectiveness, and accuracy? Is implementing artificial intelligence alongside human intelligence beneficial or a colossal mistake? Are AI's in the workforce helpful or harmful? This paper will discuss facts on the utilization of Artificial Intelligence, specifically, machine learning in work environments, and assess if overall, ML has a positive impact on the way humans and businesses function. That is, can machine learning enhance human learning? Machines have been becoming more and more prevalent in business and lives since the 1950s and 60's when IBM became the first dominant mainframe company (Bourgeois 2014). For example, the medical industry is one that uses mostly AI for day-to-day functioning. Sources claim that in the future, it is forecasted 80% of a doctor's responsibilities will be completed by some sort of machine, and the industry of medical AI will be worth $10 billion (Longoni 2019). The rate at which machine learning advance is staggering, as conceptualized in Figure 1. The figure also illustrates an industry’s leader’s certainty that as technology advances, AI and ML developments will speed up even more. At that rate of change, human learning cannot bridge the gap with machine learning. By association, for most humans, human speed of learning could not reach the level of reaction and recall time that machines have by virtue the machines’ computation time, which is on the order of milliseconds. As will be discussed later in this paper, there are ways to enhance human learning by coupling humans with machines. If not, human jobs and entire careers could potentially be lost (Bolton 2018). The authors will present a framework to bridge the learning gap and supplement human learning with cues that will allow humans to enhance their learning recall. While it's unlikely that computers will be able to replace humans in the most complex, nonstandard, or most creative tasks, job replacement is a fair concern. AI is already used to replace or augment human efforts in some routine or predictable tasks, reducing the need for human labor or necessitating shifts in human tasks (Daugherty and Wilson, 2019). In addition, such use of AI might create barriers to communication between humans and machines as machines communicate with each other (Michaels 2017). While Charles Babbage created the computer as a means of calculation, it was his translator Ada Lovelace who envisioned the new machine as a creative device, able to compose music. About one hundred years later, Alan Turing posited that a computer might be capable of original ideas, contradicting Lovelace's belief on the topic. Since Turing's work in the 1950s, machines have indeed been used to create associations and rules for a given input dataset. In the twenty-first century, however, artificial intelligence and computer-aided learning are still mostly viewed as tools for facilitating collaboration, rather than collaborators with humans. Research Questions The main research questions are: (1) Can machine learning enhance human learning in work environments? (2) Can humans and machines work and learn collaboratively on a given task or assignment---rather than just being a tool for collaboration? And, if yes, how can machine learning enhance learning in times of disruption and beyond? Using a Problem-EvidenceAnalysis-Solution (or PEAS) framework, the researchers set out to state the complex problem at the intersection of machine learning and human learning. For instance, what are the issues related to the new AI technologies, and how can humans benefit from using the available ML technologies. Research on skill-biased organizational change (SBOC) has shown that in times of rapid technological growth, the jobs replaced are those who require routine tasks. Technology tends not to decrease the need for employees who make complex decisions and those who address nonroutine tasks. However, so far, the research has focused primarily on job replacement, rather than how ML can be used to increase the skill of workers. It is essential to understand the role of ML in job enhancement, because this may provide the key to employee retention in times of disruption (Autor, D. H., Levy, F., & Murnane, R. J. 2003; Caroli, E., & Van Reenen, J. 2001). Problem Although given a negative connotation, the term "disruption" can refer to any rapid change in employment. Of course, jobs can be affected during societal and business disruptions by external environmental factors, such as AI, automation, famine, wars, hurricanes, and pandemics. But disruptive change can also be caused by internal environmental factors in businesses, such as, product obsolescence, or not adapting quickly to technological innovations, such as machine learning. Humans may find it extremely challenging to adapt to new situations and technologies since the new work environment requires specialized skills and knowledge in order to transition to new job opportunities quickly. But rather than AI becoming a threat or problem, machines can play a supportive role why individuals gain new technical knowledge. Society is impacted both economically and socially when jobs are scarce, due to any of the above factors. Some researchers fear that humans may eventually be superseded by AI in the workplace (Bolton, et al. 2018). AI is already present in daily life---in smartphones, chatbots, Grammarly, and Alexa, for instance. And humans benefit from its conveniences. However, there is uncertainty about future applications of AI which engenders worry about job replacement (Andrews 2018). Are all these innovations making us smarter or more dependent? Evidence How do humans learn? This paper focus on how ML can be used to provide supplemental learning cues to enhance learning without impinging on the natural learning process. As one of the authors commented "it will be almost like having a spouse or mentor providing cues when forgetting a family adventure or anecdote" (Personal Communication, 2020). Researchers have explained how providing environmental simulations in the work environment (i.e., variety of colors, scenery, and sounds) while humans try to learn a subject, can enhance recall (Carey, 2014). Similarly, multi-sensorial situations or multimedia experiences can be provided using ML-triggered sensory simulations in a timely fashion. Nevertheless, before venturing into the potential applications of artificial intelligence (AI) and machine learning (ML), it is useful to establish an understanding of human learning. By experimenting and reviewing the research learning literature, investigators have discovered the best ways to learn most subjects—explicitly and implicitly. Human learning occurs not only through rote memorization and experimentation but also through emotional and experiential connection with the subject matter. There is evidence that perceived competence at a task incentivizes learners to advance in ability (Covington, 1984). Complexity is also related to the desire to learn through the invocation of feelings of arousal and anxiety (Hamari et al, 2016). Other emotions such as the feeling of immersion and enjoyment contribute favorably to learning outcomes (Rosenbaum et al. 2007, Czikcentmihalyi, 1990, Newmann 1992) demonstrated that the perceived authenticity of a learning experience can facilitate the feeling of immersion for students. The neurological understanding of learning involves forming, removing, or altering neural brain connections made throughout the learning process (Bassett et al. 2017). Human learning involves two distinct processes. Humans can learn either implicitly or explicitly. Through implicit learning, humans learn first-hand by doing something or are taught either how to do something, how not to do something, or how to change how we do something. Implicit learning is much more complex, however. Through this form of learning, humans learn through their perception of the world around them. For instance, this can be by observing someone else perform a task and seeing the outcome. If humans experience failure on a specific task, it is likely they won't perform the task in the same way they did as if they perform it successfully. In this process, humans would have learned something useful from failure or success. The clear distinction between these two types of learning is that implicit learning is subconscious and explicit learning is conscious. What is Artificial Intelligence and Machine Learning? Simply put, artificial intelligence (AI) is the study of training computer systems (rather than programming computers) to accomplish the tasks that humans can perform using their cognitive abilities. AI can emulate and simulate human-like abilities and functionalities, such as "learning" from the work environment (datasets). Types of AI include reactive-machine, limited-memory, theory-of-mind, and self-awareness. Reactive-machine involves only the present information and cannot form memories. Video games are an example since they respond to different stimuli. This primitive AI is unable to learn, while limited-memory can look into the past and learn (Hintze, 2016). This category of AI comprises more sophisticated artifacts, such as chatbots and biometric authentication. Theory-of-mind and self-awareness are beyond the scope of this paper and are considered underdeveloped and, to some extent, hypothetical. Reactive-machine algorithms are best suited for tasks that are repeated and predictable. For those that require complex or changing input for which a list of if-then statements is not feasible, machine learning may be applied. Machine learning uses past outcomes to construct and change its own algorithms to minimize the difference between proposed solutions and desired outcomes, or calculated solutions and the population. Machine learning (ML) fits into the general category of AI since ML can interpret data, potentially learn from the data, and then use that acquired knowledge to adapt and achieve specific goals (Green and Green 2019). ML is the process of learning with training labels. These systems are like humans in the sense that they must learn heuristically – by iterating in problem-solving process and comparing against the given assumption. In this case, since an algorithm "learns" from a labeled dataset, it provides an answer key that it can use to evaluate the accuracy on the training data. This process is known as supervised learning or supervised machine learning. But note that humans must still label data into this ML system and monitor the learning process. ML can augment or complement human capabilities (Wilson and Daugherty, 2018, n.p.). However, AI or ML cannot fully replace the human workforce yet, given humans' tacit real-world knowledge, intuition, and experience interacting with the physical world (Emerging Technology from the arXiv 2018). Another type of ML is called reinforcement learning, or reinforcement machine learning. Video gaming is one of these types of ML where cues are provided to succeed in the target goal. In this type of ML, AI agents are trying to attain a goal or improve the effectiveness and performance of a given task. Therefore, it is an interactive process—employing trial and error to arrive at a solution of a problem. Organizations that can devote the time and resources to developing and integrating machine learning in their operations will reap the benefits of automated business functions. Companies that do not incorporate ML will be left behind (Dirican 2015). The encouraging news is that job growth is expected to offset jobs lost to automation (Lund 2017, 11). Jobs introduced by the development of machine learning will most likely be enough to offset the jobs that were taken by AI and machine learning (Lund 2017, 11). But the skillsets required from human are different. Many tasks are riddled with functions and processes that are completed by ML. Most of the benefits of ML are now commonplace - personalized ads, personalized recommendations, and optimization of websites are driven by AI and ML algorithms. The solution to improving human learning coupled with machine learning should build upon the algorithms developed for some of the above applications. It is clear that AI and machine learning have been critical in not only improving our quality of life but also improving how we work. AI can be used when it complements the creativity, emotion, and innovation of humans. There are instances when AI can threaten to replace human jobs. When AI researchers were surveyed a few years ago, the survey revealed a 50% possibility that AI will be surpassing humans in all activities in the next 45 years (Maini and Sabri 2017). Now, truck drivers and call center operators can be replaced by automation, speech recognition and deep neural network robots (Brynjolfsson, Rock, and Syverson 2017). Brynjolfsson, et al (2017) stated that AI could perform 45% of the tasks supplied by the human labor force. In times of technological change, displaced workers will possibly have to learn new skills to find different jobs. On the other hand, AI can also improve work performance and value creation. Although AI can replace some occupations, it can also bring benefit to employees and employers in the workplace environment. Evidence gathered about how humans continue to learn and create value in the workplace and how machine learning enhances on-the-job training as well as how employers' benefit from improved work performance with AI paints a different picture. Even though AI can drastically change the workplace, AI will not yet replace the human workforce, but instead can augment, support, supplement, and complement human-abilities and capabilities (Wilson and Daugherty 2018). Today, humans still learn things that machines cannot because of their background knowledge and real-world experience (Emerging Technology from the arXiv 2018). With AI freeing workers from everyday responsibilities like administrative work, employees can now focus more on strategic tasks (Wellers, Elliott, and Noga 2017). There are many human skills that AI cannot yet emulate or perform, such as social networking, staff development and coaching, and collaboration (Kolbjornsrud, Amico, and Thomas, n.d.). But AI can aid the process. More importantly, AI does not understand the potential consequences for mistakes or moral implications of decisions (Green and Green 2019). Hence, humans can continue to add value in the workplace by focusing on skills that machines are incapable of, such as creativity, collaboration, empathy, and making ethical judgments calls (Kolbjornsrud, Amico, and Thomas, n.d.). Since Machine Learning analyzes data through mathematical processes, known as algorithms, humans are still essential in reviewing this information to make smart and informed decisions (Finlay 2018). For example, AI chatbots blend past customer service data and natural language processing, along with algorithms that are continually learning through exchanges, to answer customer questions (Wellers, Elliott, and Noga 2017). These chatbots still need to learn from human workers for when exceptions occur. Humans provide the labels and meaning to AI with their experience from everyday life (Green and Green 2019). Fortunately, human time saved from routine customer service calls, workers can focus on creative strategic-planning and strengthening customer retention, for instance. Without human-supervision, AI's success is limited. Humans and AI in the workplace can work together to enhance each other's strengths and compensate for each other's weaknesses. People are needed to explain the AI system's predictions, outputs, and possible errors to others who are not familiar with AI (Green and Green 2019). Workers must also inspect algorithms for fairness rather than biased results. Humans must capitalize on traits such as their positive motives, societal values, and good decision-making skills (Green and Green 2019). Workers who embrace an opportunity to collaborate with ML can also learn from the interaction with machines and continue to add value in the workplace. For instance, when playing a game of Chess or Go with an AI machine, humans can gain unique insights not easily replicated when playing even with a master. How Can Machine Learning Enhance Human Learning (On-the-Job Training)? Research gives evidence that ML can improve on-the-job training, particularly through just-intime e-learning. Just-in-time e-learning offers knowledge to workers with teaching, learning, and training at their workplace fingertips (Lokaiczyk, et al. 2007). This ML process learns the employee's work tasks and then knows how and when to supply pinpointed training (Lokaiczyk, et al. 2007). Theories and software programs such as Advanced Process-Oriented Self-Directed Learning Environment (APOSDLE), heighten on the job training by improving worker's knowledge and abilities, ultimately increasing productivity (Lokaiczyk, et al. 2007). In addition to personalized coaching, ML provides reinforced training, measurable effectiveness, and Return on Investment (ROI) along with access to an infinite amount of information (Riebli 2018). Research shows that humans need training reinforcement (i.e., repetition) to retain the information learned; and AI can provide this environment (Riebli 2018). Through personalized reinforcement learning process, ML can provide recommendations on what an employee needs to learn next. Training procedures can be improved with AI measuring results and enabling adjustments as required, providing more cost-effective, effective, and proficient training (Riebli 2018). With ML enriching employee training and development, supervisors can allot more time to interact with employees as well as improving employee relations, retention, and innovation (Riebli 2018). Table 1 shows ways in which machine learning can enhance human learning. In short, ML enhances human learning through personalization, customization, engagement of senses, challenge, and correction. AI in the workplace automates routine tasks allowing for employees to be more creative and productive (Lee 2018). Presently, top companies are utilizing ML tools and devices for smart decision-making automation (Wellers, Elliott, Noga, 2017). Intelligent decision making directly affects company costs and revenue. Statistics show that data-driven companies are 5% more productive and 6% more profitable (McAfee and Brynjolfsson 2012). Businesses are also using AI for digital transformation (Wellers, Elliott, and Noga 2017). By 2025, AI is predicted to become a $100 billion market, drastically altering the way business is handled and impacting the workplace (Wellers, Elliott, and Noga 2017). Businesses can benefit from AI by enhancing employee work performance to add value and competitive advantage to organizations. The Society of Human Resource Management (SHRM) describes that AI technology, together with machine learning and cloud-based analytic platforms, can analyze today's employees and their work performance, enabling higher quality outcomes (Romeo 2019). Table 1: Ways in Which Machine Learning Can Enhance Human Learning Means Reason Source Likelihood Unique and efficient learning Intelligent interfaces allow for customization to meet each learner's requirements. Real-time feedback (Riebli 2018) High offered during training can fix mistakes early and prevent reinforcement of negative behaviors. Improved training retention Intelligent systems reinforce the main points of lessons. (Riebli 2018) High Quantified value & ROI The capability of monitoring training leads to adjustments as necessary. Collaboration potential increases. Human learning proficiency is improved. (Riebli 2018) High Access to a virtually infinite amount of information Big Data equates to more information and potentially better decisions. Big Data Analytics may increase productivity and profitability. Decisions can be made based on evidence, rather than "feel." (McAfee and Brynjolfsson 2012) Medium-high Analyze and diagnose human difficulties learning the material. Mistakes can be recognized immediately, and corrective behavior can be prescribed. (Authors 2020) Medium-high Make recommendations to speed up the learning process. ML may recognize aspects of a task that have already been mastered. (Authors 2020) Medium-high Sense human frustration with the material and propose alternative learning paths. ML may possess multiple workflow paths, and it may suggest alternate paths based upon the learner's responses. (Authors 2020) Medium-high Trigger simulation of a variety of geographic locations to help neurons create more connections with the material The ML software may follow learning paths that require movement of the learner. (Authors 2020) Medium-high Add music, color, or other sensorystimulating backgrounds that would simulate the test environment. ML can incorporate multiple senses to build triggers for recall. Enhance human recall. (Authors 2020) (Carey 2014) Medium-high Adjust the difficulty of the task such that the learner is challenged but not overwhelmed ML can create task difficulty that avoids instant mastery and cultivates learning. This is demonstrated in 'just-in-time elearning.' Detecting training as it is needed. (Authors 2020) (Lokaiczyk, et al. 2007). (Carey 2014) Medium-high Detect fatigue and suggest work breaks ML can suggest breaks when the learner appears to be completing tasks more slowly or with more errors. (Authors 2020) Medium-high With human supervision: "Detect inductive biases which make it possible for a ML can challenge inductive biases and target the learner to analyze differences (Griffiths et al 2008). Low-high, depending on the training dataset. learner to choose between hypotheses that are equally consistent with the observed data." between the choices. With human supervision: Compare models "that differ in their inductive biases to the performance of human participants solving an inductive problem in the laboratory, it becomes possible to evaluate which set of inductive biases most closely matches those of human learners." ML can learn where inductive biases are likely to exist, through training. (Griffiths et al 2008). Low-high, depending on the training dataset. Source: Developed by the authors with adaptions from the tabulated references. Pattern recognition and monitoring patterns can also enhance employee productivity and work performance (Romeo 2019). The planning process is an excellent example of this AI system capability. AI can aid in scheduling and planning for higher efficiency, which results in boosted employee morale, productivity, and increasing cost saving for the employer (Romeo 2019). AI and ML increase efficiency and effectiveness throughout an organization, generating a competitive advantage in the global economy (Romeo 2019). Further, machine learning is allowing businesses to streamline procedures, boosting employee morale, increasing production, and enriching customer relations (Wellers, Elliott, and Noga 2017). Analysis Even routine tasks imbue the employee with tacit knowledge that is complementary to AI. As mentioned earlier, research indicates that businesses that replace human workers with technologies, such as AI and ML, may experience short-term productivity gains (Wilson and Daugherty 2018, n.p.). The evidence presented here suggests that there may yet be untapped opportunity for government and industry to harness human capital within the organization. This might occur, for example, by equipping employees with the skills they would need to adjust to and unite with the AI interface. Further, it would allow smart organizations to capitalize on the collaboration of machine learning and human learning. Companies that embrace AI and ML while cultivating a focus on employee learning may realize significant competitive advantage. An alternative offered to those humans who fear being replaced by machines comes from the Historian Yuval Noah Harari. Harari urges humans to focus on building emotional intelligence and mental balance (Skipper 2018). If people want to stay employed in the AI era, they must continually reinvent themselves. Harari believes that the most valuable employability asset in the future will be psychological flexibility. Artificial Intelligence will continue to evolve, forcing workers into continued learning, education, and training for new jobs. Change is stressful. When specific jobs become extinct, people need adaptability to learn new skills and take on innovative jobs. Since it is human nature to resist change, humans must practice managing transition. With awareness and self-knowledge, workers can repeatedly reinvent themselves, preserving employment throughout the Artificial Intelligence Revolution (Skipper 2018). A second alternative, Dr. Kai-Fu Lee suggests, is for people to espouse their human traits. Humans possess compassion, love, and creativity that no machine can acquire. Lee believes that since AI has alleviated ordinary daily tasks, humans can now get in touch with what is truly important in life, formulating a new plan for life. Dr. Lee encourages humans to change the mentality that work defines them. Create a new life that emphasizes the importance of human values and prioritization, embracing time spent with loved ones. Lee claims that AI will replace various human jobs over the next fifteen years. Yet, new occupations can focus on humanity and helping others such as social workers, caregivers, and teachers. Lee recommends capitalizing on human brains and mortal hearts. The coexistence of AI and humans' ability to love can allow humanity to retain value in the face of more productivity-enhancing technologies (Lee 2018). Solution Allowing employees to retain jobs even as the skills change, with the help of AI/ML. After appraisal of the above analysis, the researchers suggest integrating both of Dr. Lee's suggestions where coexistence is the key function. Employer and employee synchronization, coupled with worker agility and adaptability, will allow the human workforce to endure in the face of the threat of AI replacement. And, better yet, machine learning can enhance human learning and vice versa. Collaborative intelligence exemplifies the requirement for this collective effort. Companies and workers must work together towards continuous learning and re-creation of self. Psychological flexibility with self-awareness is an essential skill needed for successful employability and continuous learning in the AI revolution. Coevolving with AI and re-creating a life built on personal values, priorities, and love is a solution to having rewarding work. Below are the components and benefits of the recommended solution. Improving Learning ML-based learning platforms should simulate the emotional, audiovisual, tactile, and experiential factors which serve as incentivizing mechanisms for students. Through feedback from the individual learner, such technology can adjust strategies "on the fly" to induce the best outcomes for that person, with that subject matter, at that time. Benedict Carey (2014) researched how humans learn as well as when, where, and why it happens. Carey found that studying at different locations (rather than the same study place) enhances learning, since subject matter will have more place and sounds associations in the learners' neural network. Researchers have gone digital to provide a variety of simulated environments incurring the costs of on-site training. AI and ML can generate new photorealistic environments to simulate medical “hot zones,” for instance. An example of the symbiosis of human and machine learning is in the field of dermatology. Convolutional neural networks can be used to identify cancerous skin lesions (Esteva et al, 2017). Training such networks relies on human classification of a known set of images. Subsequently, new images can be classified automatically. The accuracy of AI classification relies upon the inclusion of all known types of cancerous lesions, so as new types are discovered, there must be human intervention to add them to the training data set. Students may test their ability to identify skin cancers using an AI “teacher” that adjusts image difficulty to match the learning rate of the student. Improving Processes Employees in the workplace are involved in a number of routine processes. Implementing ML into routine processes can help workers devote more time to improving processes. For instance, an employee in human resources may use ML for data gathering or use it to mine customer actions and transactions. Further, social sentiment data can be used to identify customers who are at high risk of withdrawing from the business relationship. Combined with profitability data, this would allow organizations to optimize for what is called next-best-action strategies as well as to personalize the end-to-end customer experience (Wellers, 2017). In addition, using ML to improve customer service is a great way to enhance human learning. The ML can sort and organize the up-to-date data and information to help employees perform better and assist the workers learn and adapt to giving the customers the best service possible. Further, ML can rapidly identify key requirements in job candidates' applications and select candidates who have the credentials that are most likely to achieve success at the company. Of course, AI software will have to be coupled with human supervision to combat human bias that might be embedded on the dataset. And can automatically flag biased language in job descriptions or detect highly qualified candidates who might have been overlooked because they didn't fit cultural expectations (Wellers, 2017). When recruiting job applicants, ML can again speed up the review process and eliminate unqualified workers. Routine processes performed by human can be tiring and/or confusing. While a recruiter can quickly scan and read multiple applications, ML never gets tired and can supply important cues to the human reviewer. Since ML can monitor existing processes and learn to recognize patterns and context, it can significantly increase the number of documents (i.e., invoices) that can be matched automatically. This would help reduce the amount of work outsourced to global service centers and, more importantly, allows the employees to focus on strategic tasks (Wellers 2017). Managing finances can be difficult for most workers due to how it can be a difficult and/or confusing task to deal with. By using ML, companies can have software set up and organize the company's finances in a way to help workers learn more efficient ways to organize it themselves. Beyond the limited scope of this paper, it should be noted that advanced research is needed on Rademacher Complexity, which measures the ability to assimilate data into a known context or form a pattern with it. While this may support the notion that a high Rademacher Complexity indicates a facility with learning, it also indicates the possibility of overfitting the data (Zhu et al. 2009). In addition, ML can help staffers in the implementation of incremental business tasks, by analyzing datasets and evaluating current systems and platforms (Craven 2019). Using specific software to aid in managing resources can help businesses with resource and data management. Further, workers can learn from ML in their workplace by managing inventory better and, in general, improving budgets allocation. Conclusion This research has illustrated how machine learning can enhance human learning in work environments. When companies recognize employees' fear of being replaced by AI, employers must provide incentives for employee engagement with machine learning to alleviate anxiety and ensure success (Wellers, Elliott and Noga, 2017). Research points that embracing AI technology in the workplace can even result in the production of new jobs (Bolton, et al. 2018). Through creating innovative, functional work relationships between AI and humans, workers can be more productive, allowing companies to capitalize on the wave of AI (Bolton, et al. 2018). The research presented here is limited to a theoretical discussion of ML and human learning. It could be further enhanced by an experimental study that compares job accuracy between subjects who train with ML and those who do not. We also do not attempt to hypothesize differences by industry. This is another avenue for future research. Not only may ML have industry-differentiated effects, but there may be contexts in which machine learning is not desirable. For instance, the market for hand-crafted goods such as quilts, knitted items, hand-built tables, and other antique-store fare signals the value of traditional familial learning methods. On the other hand, the growth of craft breweries in the 2000s indicates the increase in consumer preference for small-batch items. In both cases, uniqueness is valued. Machine learning may be a more valuable asset in the latter case, in which taste regional preferences can be used to optimize flavor profiles for local breweries. As analyzed, explained, and tabulated in this paper, machine learning can enhance human learning in multiple ways—namely: (1) ML intelligent interfaces can easily allow for personalization and customization to meet each learner's requirements; (2) ML can offer realtime feedback during training and can help fix mistakes early and prevent reinforcement of negative behaviors; (3) ML intelligent systems can reinforce main points of lessons; (4) ML can provide capability for monitoring training leads and adapt training as necessary (5) Human learning and ML collaboration can increase, as datasets are nourished; (6) human learning proficiency can improve from the collaborative interaction with ML; (7) access to big data repositories signify more information and potentially better decisions; (8) Big Data analytics can increase productivity and profitability; (9) arrangements can be made collaboratively based on both objective evidence and subjective human abilities, such as ethical and social impacts; (10) mistakes can be recognized immediately, and the corrective behavior can be prescribed; (11) ML can recognize aspects of a task that have already been mastered and help learners move to the next assignment; (12) since ML may possess multiple workflow paths, it can suggest alternate routes based on the learner's responses; (13) The ML system can follow learning paths that require specific location and movement of the learner; (14) ML can trigger simulations of a variety of geographic locations to help neurons create more connections with the material; (15) ML speech emotion recognition systems can sense human frustration with the content and propose alternative learning paths; (16) ML can create task difficulty that avoid instant mastery and cultivate challenging learning, as demonstrated in just-in-time e-learning, by detecting training as needed; (17) ML can incorporate multiple senses to build triggers for recall and, therefore, enhance and reinforce lessons learnt; (18) ML can add music, color or other sensory stimulating backgrounds that would simulate a given work or test environment; (19) ML can detect the learners' fatigue and suggest work-breaks, or even suggest breaks when the learner appears to be completing tasks more slowly or with more errors; (20) with human supervision, ML can detect inductive biases. For instance, ML can allow users to choose between hypotheses that are equally consistent with the observed data. Further, ML can challenge inductive biases. And cue the worker or learner to analyze differences between the available choices, and (21) ML can offer real-time feedback during training to fix mistakes early and prevent reinforcement of negative behaviors on both workers and learners, among other functionality. As in the previous technological and industrial revolutions, the AI revolution will unlock new opportunities for human growth and development, paving the way for job creation in areas previously unknown. ML will not be substituting humans in the foreseeable future since creativity and persuasion are not yet within the realm of possibilities. Nevertheless, humans and machines can undoubtedly collaborate and together navigate an uncertain future in times of disruption and beyond—either as useful tools or as virtual collaborators in the near future. REFERENCES Autor, D. H., Levy, F., & Murnane, R. J. (2003). The skill content of recent technological change: An empirical exploration. The Quarterly journal of economics, 118(4), 1279- 1333. Andrews, Whit. 2018. "The Real Truth of Artificial Intelligence." AI Gartner Webinar.pdf. file:///E:/ISM%203011/AI%20Report/AI%20Gartner%20Webinar.pdf Bassett, Danielle S., and Marcelo G. Mattar. "A Network Neuroscience of Human Learning: Potential to Inform Quantitative Theories of Brain and Behavior." Trends in Cognitive Sciences. Elsevier Current Trends, March 2, 2017. https://www.sciencedirect.com/science/article/pii/S1364661317300165. Bourgeois, David. 2014. Information Systems for Business and Beyond. Open Textbook Library. https://open.umn.edu/opentextbooks/textbooks/information-systems-forbusiness-and-beyond Bolton, Charlynne, Veronika Machová, Maria Kovacova, and Katarina Valaskova. 2018. "The Power of Human–Machine Collaboration: Artificial Intelligence, Business Automation, and the Smart Economy." Economics, Management and Financial Markets 13(4): 51– 56. doi: http://dx.doi.org/10.22381/EMFM13420184. Brynjolfsson, Erik, Daniel Rock, and Chad Syverson. 2017. "Artificial Intelligence and the Modern Productivity Paradox: A Clash of Expectations and Statistics." The National Bureau of Economic Research. Nov. 2017. https://www.nber.org/papers/w24001 Caroli, E., & Van Reenen, J. (2001). Skill-biased organizational change? Evidence from a panel of British and French establishments. The Quarterly Journal of Economics, 116(4), 1449-1492. Covington, M. V. (1984). The self-worth theory of achievement motivation: Findings and implications. The elementary school journal, 85(1), 5-20. Craven V. (2019, September). How Machine Learning Can Help with Energy Management: Artificial Intelligence and its Impact on Building Optimization, Leasing and Service. Retrieved from Buildings, Sep2019, Vol. 113 Issue 8, p 18-21. 3P Carey, Benedict. 2014. How We Learn: The Surprising Truth About When, Where, and Why It Happens. Random House. New York. Daugherty, Paul and H. James Wilson. 2019. "Using AI to Make Knowledge Workers More Effective." Harvard Business Review, April 19 2019, accessed July 14, 2020. https://hbr.org/2019/04/using-ai-to-make-knowledge-workers-more-effective. Dirican, Cüneyt. 2015. "The Impacts of Robotics, Artificial Intelligence on Business and Economics." Procedia Social and Behavioral Sciences 564-573. https://cyberleninka.org/article/n/926024.pdf. Emerging Technology from the arXiv. 2018. "Why humans learn faster than AI-for now." MIT Technology Review. Mar 7, 2018. https://www.technologyreview.com/s/610434/whyhumans-learn-faster-than-ai-for-now/ Esteva, Andre, Brett Kuprel, Roberto Novoa, Justin Ko, Susan Swetter, Helen Blau, and Sebastian Thrun. 2017. “Dermatologist-level classification of skin cancer with deep neural networks.” Nature, 542, 115-118. Finlay, Steven. 2018. Artificial Intelligence and Machine Learning for Business A No-Nonsense Guide to Data Driven Technologies. 3rd ed. Great Britain: Relativistic. eBook. http://file:///E:/ISM%203011/AI%20Report/AI-and-Machine-Learning-for-BusinessThird-Edition- Sample.pdf. Friedman, T. L. 2016. Thank You for Being Late: An Optimist's Guide to Thriving in the Age of Accelerations. New York, NY: Farrar, Straus and Giroux. Giamellaro, M. (2014). Primary contextualization of science learning through immersion in content-rich settings. International Journal of Science Education, 36(17), 2848-2871. Green, John and Hank Green. 2019. "Humans and AI Working Together." #14. PBS. November 22, 2019. Video, 9:56. https://www.pbs.org/video/humans-and-ai-working-together-14- lathei/ Green, John and Hank Green. 2020. "The Future of Artificial Intelligence." #20. PBS. January 2, 2020. Video, 10:50. https://www.pbs.org/video/the-future-of-artificial-intelligence-20- pbq7fm/ Griffiths, T. L., Christian, B. R., & Kalish, M. L. (2008). Using category structures to test iterated learning as a method for identifying inductive biases. Cognitive Science, 32(1), 68-107. Haenlein, Michael, and Andreas Kaplan. 2019. "A Brief History of Artificial Intelligence: On the Past, Present, and Future of Artificial Intelligence." California Management Review 61 (4): 5–14. doi:10.1177/0008125619864925. Hamari, J., Shernoff, D. J., Rowe, E., Coller, B., Asbell-Clarke, J., & Edwards, T. (2016). Challenging games help students learn: An empirical study on engagement, flow and immersion in game-based learning. Computers in human behavior, 54, 170-179. Hintze, Arend. "Understanding the Four Types of Artificial Intelligence." Government Technology. Last modified November 14, 2016. Accessed March 27, 2020. https://www.govtech.com/computing/Understanding-the-Four-Types-of-ArtificialIntelligence.html . Kolbjornsrud, Vegard, Richard Amico, and Robert Thomas. n.d. "AI Could Liberate 50% of Managers' Time." Harvard Business Review. Accessed March 21, 2020. Video. https://hbr.org/2017/05/8-ways-machine-learning-is-improving-companies-workprocesses Lee, Kai-Fu. 2018. "How AI can save our humanity." TED. August 27, 2018. Video, 14:49. https://www.youtube.com/watch?v=ajGgd9Ld-Wc Lokaiczyk, Robert, Andreas Faatz, Arne Beckhaus, and Manuel Goertz. 2007. "Enhancing Justin-Time E-Learning Through Machine Learning on Desktop Context Sensors." Modeling and Using Context. Roskilde, Denmark: Springer. August 20-24, 2007. https://link.springer.com/content/pdf/10.1007%2F978-3-540-74255-5_25.pdf Longoni, Chiara, Andrea Bonezzi, and Carey K Morewedge. 2019. "Resistance to Medical Artificial Intelligence." Journal of Consumer Research 46 (4): 629–50. doi:10.1093/jcr/ucz013. Lund, Susan, Michael Chui, Jacques Bughin, Jonathan Woetzel, Parul Batra, Ryan Ko, and Saurabh Sanghvi. 2017. [Jobs lost, jobs gained: What the future of work will mean for jobs, skills, and wages. Study, San Francisco: McKinsey Global Institute. https://www.mckinsey.com/global-themes/future-of-organizations-andwork/what-thefuture-of-work-will-mean-for-jobs-skills-and-wages. Maini, Vishal and Samer Sabri. 2017. "Machine Learning for Humans." August 19, 2017. PDF. file:///E:/ISM%203011/AI%20Report/Machine%20Learning%20for%20Humans.pdf McAfee, Andrew and Erick Brynjolfsson. 2012. "Big Data: The Management Revolution." Harvard Business Review. PDF. file:///E:/ISM%203011/BIG%20DATA%20Harvard%20Review.pdf Michaels, Vivian. 2017. The Unforeseen Consequences of an AI-built Social Network." September 15, 2017. VentureBeat. https://venturebeat.com/2017/09/15/theunforeseen-consequences-of-an-ai-built-social-network/ Rodney, Harriet, Katarina Valaskova, and Pavol Durana. 2019. "The Artificial Intelligence Recruitment Process: How Technological Advancements Have Reshaped Job Application and Selection Practices." Psychosociological Issues in Human Resource Management 7 (1): 42–47. doi:10.22381/PIHRM7120194. Riebli, Jennifer. 2018. "3 Ways Artificial Intelligence Can Improve Workplace Learning." Jan. 25, 2018. https://bthechange.com/3-ways-artificial-intelligence-can-improve-workplacelearning-40b1185bc0d3 Romeo, Jim. 2019. "Artificial Intelligence Boosts Employee Performance." SHRM Better Workplaces Better World. September 13, 2019. https://www.shrm.org/resourcesandtools/hr-topics/technology/pages/artificialintelligence-boosts-employee-performance.aspx Rosenbaum, E., Klopfer, E. & Perry, J. On Location Learning: Authentic Applied Science with Networked Augmented Realities. J Sci Educ Technol 16, 31–45 (2007). https://doi.org/10.1007/s10956-006-9036-0 Skipper, Clay. 2018. "The Most Important Survival Skill for the Next 50 Years Isn't What You Think." GQ. https://www.gq.com/story/yuval-noah-harari-tech-future-survival Wellers, Dan, Timo Elliott, and Markus Noga. 2017. "8 Ways Machine Learning Is Improving Companies' Work Processes." Harvard Business Review. May 31, 2017. https://hbr.org/2017/05/8-ways-machine-learning-is-improving-companies-work-processes Wilson, James, and Paul Daugherty. 2018. "Collaborative Intelligence: Humans and AI are Joining Forces." Harvard Business Review. July-August 2018. https://hbr.org/2018/07/collaborative-intelligence-humans-and-ai-are-joiningforces?referral=03758&cm_vc=rr_item_page.top_right Zhu, J., Gibson, B. R., & Rogers, T. T. (2009). Human Rademacher complexity. In Advances in neural information processing systems (pp. 2322-2330). ABOUT THE AUTHORS Walter Rodriguez: Founder and Chief Learning Officer at Coursewell.com (Adaptiva Corp) in Bonita Springs, Florida & Naples, Florida. Formerly, Professor of Information Systems at the Lutgert College of Business and Founding Director, Institute for Technological Innovation, Florida Gulf Coast University, Fort Myers, Florida, USA. Patricia Angle: Assistant Professor of Management Information Systems, University of Oklahoma, Norman, Oklahoma, USA. Formerly, Assistant Professor of Information Systems at the Lutgert College of Business, Florida Gulf Coast University, Fort Myers, Florida, USA. Michele Snyder: Undergraduate Student, Lutgert College of Business, Florida Gulf Coast