How Good Is AI at Improving Training?

Understanding how artificial intelligence works is the first step in figuring out how we can apply it to training.

When I think of artificial intelligence (AI), I can’t help but jump back to my first exposure to the concept with the film, 2001: A Space Odyssey. Called HAL 9000, the computer in the movie runs by artificial intelligence and is the main antagonist in Arthur C. Clarke’s “Space Odyssey” book series.

HAL initially is portrayed as a dependable member of the crew, maintaining ship functions and engaging genially with its human crewmates on an equal footing. Then malfunctions occur in the system and the astronauts try to shut it down. However, HAL takes over the spaceship’s system in order to fulfill its programmed directives and ends up killing one of the astronauts.

Thank goodness, this is just science fiction—it’s not exactly what you want happening with today’s workplace training.

But we are more used to machine learning than we realize. Take, for example, the Internet searches we conduct, working with virtual assistants, or even shopping online and checking out recommended purchases.

Interestingly, in a recent survey of HR professionals by the Engage2Excel Group, when asked which technology trends HR professionals were most interested in learning about, 81 percent chose AI.


Let’s dive into the beginning steps toward AI. Rule-based machine learning (RBML) is a term computer science uses to include any machine learning method that identifies, learns, or evolves “rules” to then store, manipulate, or apply those rules.

Data fed into a computer can handle multidimensional and multi-varied inputs. No human intervention is required, and computer analysis can easily identify various trends and patterns. This has wide applications, including training, and allows organizations to use data synthesis and analysis for ongoing continuous improvement.

The disadvantages of machine learning include acquiring the right kind of data, which consumes a fair amount of time and people to access. This human element of data selection can generate a susceptibility to higher error levels, as well as questionable interpretation of data.

SAS, a leader in business analytics software and services, states that machine learning tackles tasks in four primary ways through:

  • Machines that need to be taught by example before they can apply the resulting insight to similar tasks.
  • Machines that can extrapolate from a general pattern and apply it to other data.
  • Machines that can, unsupervised, study data to find patterns, getting better with experience (though never autonomous).
  • Machines that can work with and exploit a given set of rules to move toward a desired outcome.


Deep learning is part of a broader family of machine learning methods based on artificial neural networks with representation learning. They are called neural networks because they loosely mimic the interconnected structure of the human brain to provide multilayered functionality. In this way, learning can be supervised, semi-supervised, or unsupervised.

Deep learning is only possible when AI systems have large amounts of data. Using self-improving neural networks, AI takes advantage of complex pattern- spotting through recognizing speech or images. Apple’s Siri and Amazon’s Alexa show this and help make AI simpler and more human-like.

School students using educational software Bakpax find its computer vision system capabilities automatically convert student handwriting into text. Bakpax’s AI capabilities also interpret what the student meant to say, and automatically learn how to grade student responses and how to appropriately compare them against all other student results.


When AI is applied to various learning tools and programs available to us today, Learning and Development (L&D) professionals can fully utilize these solutions to deliver learning that is:

  • Automated in its delivery
  • Personalized to the individual learner
  • Intuitive in its application
  • Data-driven across the organization
  • Fostered continuously throughout the organization

The data gained from employee usage of AI-driven learning programs will help L&D professionals better understand the insights of learners’ behavior. This helps improve the overall learning experience by developing intuitive learning pathways. AI provides a predictive analytics component that examines data, or content, to answer the question: “What is likely to happen?”

Learning content becomes much smarter in its application because it is more intuitive and responsive to learners’ needs. Learning content can be delivered through a variety of media, such as chatbots, and help with retaining the knowledge learned. This can be through ongoing delivery of courses in progress or with just-in-time learning of new material the AI system recommends for the learner.


Using AI in training will enable employees to receive learning content prescribed from knowledge and skill-based assessment tools and personalized course recommendations. Learners can receive feedback on public speaking or presentation skills, for example, because AI and machine learning can use speech-to-text processes to analyze and critique. AI also helps in making all types of learning content accessible to individuals with most forms of disabilities. Microsoft’s Seeing AI app, for example, narrates to low-vision individuals the world going on around them.

The completion rate for online training programs is typically less than 15 percent. However, with AI providing content in the learner’s preferred delivery medium and following up with chatbot reinforcer messages and links, completion rates can drastically improve.

But the biggest asset of AI and training comes back to data. Analyzing all of the training data through AI will help to provide the best kind of training for the future, delivered the preferred way, at the level of difficulty best suited to each employee.

Roy Saunderson, MA, CRP, is author of “Practicing Recognition” and Chief Learning Officer at Rideau Recognition Solutions. His consulting and learning skills focus on helping companies “give real recognition the right way wherever they are.” For recognition insights, visit: For more information, e-mail him at or visit

Roy Saunderson, MA, CRP
Roy Saunderson, MA, CRP, is author of “Practicing Recognition” and Chief Learning Officer at Rideau Recognition Solutions. His consulting and learning skills focus on helping companies “give real recognition the right way wherever they are.” For recognition insights, visit: For more information, e-mail him at: or visit: