Ivan Au; Carlos Liu; and Maria Obaid
Machine vs Deep Learning
There are two sub-sets of AI: machine learning and deep learning. The following table summarizes some of the key differences and introduces you to key terms that will be used throughout the following chapters:
| Machine Learning | Deep Learning |
| A subset of AI. | A subset of AI and machine learning. |
| Uses thousands of data points. | Uses millions of data points
|
| Identifies relationships, uses predictive analytics, and makes decisions. | Uses training data to predict patterns when new data are revealed. |
AI can be used in the workplace to improve and enhance learning outcomes when training and developing employees. Through the integration of machine and deep learning, AI and other technologies can predict patterns and enhance the learning experience of employees. It can also be used to eliminate routine tasks and leave more room for more creative tasks. Therefore, the following section will introduce you to eLearning and how employees use it for on-the-job training, the differences in learning styles and using technology, and how AI training impacts different demographics in the workplace.
eLearning
Tai (2008) describes eLearning as instruction and training provided by a teacher or self-paced learning using a program database housed on a local area network, and it is the type of AI that most organizations use in training. It can lead to better performance, quicker time to market, improved operational productivity, increased retention, and a higher return on investment (Tai, 2008).
eLearning is an essential component of training programs for many organizations that work towards achieving their strategic goals and gaining a competitive advantage.
What are the benefits of launching e-Learning?
There are various benefits for organizations that use e-Learning for job training. The following table summarizes some of those key benefits (Hartley, 2001):
| Availability | When and where the learner has access to the organization’s e-Learning applications. |
| Learner control | Learners can access only the information they require at the time they require it. They can practice skills until they master them without the rest of the class seeing them make mistakes. |
| Object-level learning | Most of the learners want a direct, object-based approach to learning since the information and knowledge they need might be buried in a large amount of content. Organizations switch the training from time-bound, event-based learning to just-in-time, process-based learning that can help learners focus on the information they need. |
| Relevant information | Given their job function, time in service, business unit, and other factors, portal technology allows information and knowledge to be delivered to people just before they need it. This trend toward just-in-time information access could force human resource development (HRD) professionals in charge of an organization’s HRD data to fundamentally alter how people obtain information. |
| Efficiency | Technology allows people to complete tasks more quickly than they could without it. For example, when using paper and pencil, most people are not as efficient as when using a calculator. |
| Reinforcement | People are more likely to use technology in a similar situation if they use it to complete a task correctly on a regular basis. |
| Immediate feedback | When a user uses technology to complete a task, the user frequently receives immediate feedback on the task’s success. When someone orders food on Uber Eats, for example, they receive confirmation immediately. |
| Appeal | Graphics, animations, and voice recordings can be revisited by learners. |
| Reduced cost | Technology transactions are normally cheaper than people transactions in e-Learning since creating e-Learning is a one-time cost, while people transactions are a repeating cost for the organization. |
| Easy to find | People are becoming more comfortable with web browsers and other related technology, so e-Learning and learning objects are becoming more readily available. |
| Reduced time in training | e-Learning applications that are well-designed can reduce traditional classroom time by as much as two-thirds since all the training material in online which reduces training time. |
How can e-Learning be used for job training?
According to Allen (2016), organizations typically employee one of the following tactics when implementing instructional strategies:
- they provide information so participants can follow instructions,
- they provide more information so that instructions can be remembered, and
- they provide even more instructions to help trainees perform at an expert level.
Misaligned Instructional Strategies (Allen, 2016)
The reason for launching eLearning is to increase the productivity of individuals to maximize the net profit. Therefore, management in organizations often see providing large amounts of information as being key to success, as shown in the graphic above. However, this is a misguided concept since information overload will inhibit trainees from reaching the desired outcome (Allen, 2016).
The three primary outcomes of eLearning are: preparing learners to follow instructions, preparing learners to perform without real-time aids, and helping learners become expert performers (Allen, 2016). For learners to follow instructions, providing information is an appropriate instructional strategy but not for other outcomes since learners need to understand the background to follow the instructions (Allen, 2016). Learners will need to practice on the job to better retain the information and fully develop their skills. Lastly, expertise is best developed through challenges with simulated outcomes as they allow trainees to adapt to unfamiliar situations with ease (Allen, 2016).
Adapting e-Learning strategies to primary performance outcomes (Allen, 2016)
If the task for learners is simple and safe, providing information will play a larger role in the e-Learning process compared to simulation (Allen, 2016). In contrast, guided practice and simulation are important for complex or potentially dangerous tasks for learners (Allen, 2016). For example, cashiers who face simple and safe tasks in their duties will benefit from being given more information at the beginning of learning and less simulation time since mistakes are harmless in this position. In contrast, surgeons who face more complex and riskier tasks on a daily basis, need more time for guided practice and simulation to minimize the risk to their patients.
What can undermine the effectiveness of e-Learning?
Tai (2008) suggests that the learner, learning material, and organizational level can undermine the effectiveness of eLearning. Learners who do not understand the purpose of the training can be poorly equipped and unmotivated. If the training content is not relevant or useful, learners will not be engaged (Tai, 2008). On an organizational level, the lack of a corporate structure or a clear strategy can impact the effectiveness of e-Learning.
How can the effectiveness of e-Learning be measured?
Kirkpatrick’s model defines how to measure effectiveness (Kirkpatrick & Kirkpatrick, 2016):
- Level 1: The effectiveness perceived by a trainee; is determined through a survey on learner satisfaction.
- Level 2: The effectiveness is measured through a learning evaluation; an examination of the content is given to make sure the learner has mastered the content of the training module.
- Level 3: The effectiveness of the training is reflected in performance improvement; learners are observed to see how their knowledge is translated into workplace performance.
- Level 4: The effectiveness of the training is determined by measuring its business impact; specifically the ROI.
