A World Economic Forum article says that artificial intelligence (AI) has affected learning and development (L&D) in two ways: First, how internal L&D teams develop content, and secondly, how the end users (employees) consume the learning. “The most significant impact is the efficiency of content creators,” says John Blackmon, chief technology officer of ELB Learning. “Many of the tools used by instructional designers are adopting AI in ways that make it faster and easier to outline content, create questions, repurpose content, create role-plays and add voiceovers.” Blackmon explains that prior to the sophistication of these tools all of these processes took time, energy and resources.

Though AI can enable improved design, development and training delivery — there’s still a limit to AI-based tools’ effectiveness. “AI can be powerful for streamlining processes, but it should not be the final decision-maker for projects that directly affect employees,” Gretchen Jacobi, head of enterprise for General Assembly, says. “AI is still prone to replicating bias and presenting hallucinations. Also … an AI-powered tool could never fully understand the idiosyncrasies of an organization at the same level as a professional development leader.”

It’s imperative that L&D professionals recognize where AI can fall short in designing training experiences, so they can step up where applicable and deliver quality learning to their workforce that leaves a noticeable impact on the business.

Using the Analysis, Design, Development, Implementation, Evaluation (ADDIE) Model, a five-phase process to creating instructional course materials and training programs, we’ll evaluate when and how to efficaciously utilize AI when designing and developing training programs, and when a human touch is needed to deliver top-notch learning content that engages learners and wins stakeholder buy-in.

How and When to use AI in Training Design: The ADDIE Model

AI is increasingly being integrated into various aspects of L&D, including content creation and training delivery. Yet, many learning leaders are challenged with knowing when AI is most effective to use in learning design and the parameters around using it within that phase.

Let’s review each phase of the ADDIE Model and how to effectively utilize AI in that process.

Analysis.

In this stage, instructional designers gather data on the business’s current learning needs, objectives, target audience, existing resources and constraints. This analysis can help identify current skills gaps and what learning is needed to fill those gaps.

AI tools can assist by:

• Analyzing the data from surveys, learning assessments and other analytics to identify training and development needs. “AI tools can help learning leaders effectively review current organizational processes and inefficiencies, and can quickly gather feedback and insights from teams in large and smaller organizations,” says Jacobi.
• Developing learning objectives from the data input to help learning leaders create more targeted and effective training content. “AI can help define learning objectives by analyzing data and metrics set by organizations,” Blackmon shares.
• Recognizing patterns in the data so learning leaders can predict learning problems and identify specific learner needs.

While AI can enhance the process of gathering, analyzing and interpreting data, ultimately, the learning leader must review the data to make a decision best for their organization and learners.

Design.

In the design phase, learning leaders use the insights gathered from the analysis to outline the instructional strategies, content, activities, assessments and delivery methods. Learning leaders use this phase to create a “blueprint” of the learning experience, including the structure, flow and sequence of learning materials.

In this process, AI can assist in:

• Drafting outlines and lists with Generative AI (Gen AI) tools like Bard, ChatGPT and Gemini. Jacobi shares that, “Gen AI tools can outline the design of straightforward training curricula and programs. That said, it’s crucial to have a human guiding the generative AI tool with effective prompting and context-specific review of the output. Gen AI tools, such as ChatGPT, can quickly create a design plan, [but] it is important for the learning leader to review and add to the outline to incorporate organization-specific elements into the design.”
• Curating content by sifting through the vast amount of data to gather relevant content, resources and examples for instructional designers to incorporate into their training materials.
• Storyboarding the learning content and recommending content layouts, templates and design ideas based on best practices and learner expectations.

AI can assist learning leaders in planning the creation of the learning material by proposing structures and sequences, and drafting an outline of the training program based on learning objectives and analyzed data. This can lend learning leaders more time to be creative and innovative in their lesson planning, however, they must still review, edit and add to the outline.

Development.

The development phase involves creating or developing the learning content based on the outlined design specifications. This may involve writing the training content, creating eye-catching visuals, developing engaging multimedia elements and building interactive components.

In this stage, learning leaders can use AI to:

• Automate tasks. “AI can automate and streamline the creation of learning materials,” says Blackmon. “For example, a trainer can input the key concepts, learning objectives and [a] preferred format into the AI-powered tools, and in a matter of minutes, it will generate content, such as quizzes and interactive elements, tailored to the course objectives.”
• Generate written content. Learning leaders can input data into natural language generation (NLG) algorithms to produce written content based on the learning objectives, subject matter expertise and instructional preferences.
• Generate visuals and design elements. And AI-driven design platforms can generate visuals and layouts for learning content, like infographics, modules and eLearning.

AI-tools have many advantages in content development, however, according to Jacobi, “Training development has the most risk for the use of Gen AI to go awry.” It’s imperative for learning leaders to check all AI-generated material for accuracy and relevancy.

“Generative AI tools can quickly assemble relevant materials (e.g., design plans, resources, images and charts) to develop a training program, however, L&D leaders must know and trust the source … there are numerous examples of these tools creating inaccurate resources or improperly citing source material.”

L&D leaders must also check the content for biasness and to ensure that it’s culturally competent and respectful. “This is a big part of the human touch … AI systems can unintentionally perpetuate existing biases present in their training data,” says Blackmon. “Learning leaders must actively monitor their output for potential biases and take steps to mitigate them.”

Learning leaders must remember that they are still accountable for the content’s quality.

Implementation.

This phase focuses on facilitating training deployment and delivery, and ensuring that the learning materials are accessible to everyone in the organization. In this process of the model, learners are receiving the learning and instruction.

AI can help learning leaders with:

• Automated feedback and assessments. AI can support this phase by grading assignments, giving learners detailed feedback and identifying areas for improvement. This can free up more time for learning leaders to focus on more personalized interactions with their learners. Blackmon shares that “AI can provide learners with instant feedback from complex NLG answers, and offer additional resources when necessary.”
• AI-powered chatbots and virtual assistants. Chatbots can deliver learning content to learners through different channels like web browsers, mobile apps or messaging platforms to bring personalized, real-time support at the time of need.

Instructional designers can utilize AI in this stage of the model to enhance the implementation of training, increasing engagement and retention and improving learning outcomes.

However, this doesn’t mean learning leaders must rely on AI alone to assess and deliver personalized support. “AI should be used to complement and enhance existing L&D processes rather than replace them entirely,” says Blackmon.

While AI may not be able to answer all questions, AI-powered chatbots can lessen the amount of questions a learning leader may receive per day. However, AI doesn’t take away their responsibility of supporting learners in their time of need. Instead, it lends them more time and energy to focus on more complex problem-solving.

Evaluation.

Evaluation is the final stage of the instructional design process, and involves assessing the effectiveness of the instructional materials and the overall learning experience to determine whether the learning objectives have been met. AI can help learning leaders collect this data so they can prove training’s impact to stakeholders, as well as continuously refine their programs.

In this final stage, AI can be leveraged for:

• Sentiment analysis. AI-powered sentimental analysis tools can analyze data from learner surveys, assessments and feedback forms to identify learners’ reactions to the content. “Gen AI tools can easily draft post-training assessments or surveys, and can aggregate responses to quickly summarize employees’ experiences before, during and after the training,” says Jacobi.
• Learning analytics. AI-powered tools can also be used to analyze and interpret learner data on how learners interacted with the training to gain insight into trends, preferences and patterns.
• Performance monitoring. AI algorithms can monitor learners’ performance post training by connecting with performance management systems, workflow automation tools or sensors embedded in the workplace equipment. “For example, an AI coach can listen to a learner’s sales pitch and detect if they are mentioning the required keywords or monitor if the learner says the words “umm” or “like” too often,” explains Blackmon. “These insights can help identify which parts of the training are most effective and which need improvement.”

AI can streamline the evaluation process by automating data collection, analysis and reporting so instructional designers can make better informed decisions to improve training and produce outcomes.

Yet, learning leaders must still be knowledgeable and skilled in data analysis and interpreting data to make effective business decisions. The AI provides the data, but it’s up to the learning leader to know how to apply that data effectively. Blackmon recommends that, “Leaders prioritize educating and training L&D staff on how to use AI tools effectively and ensure that the data used to train AI models is accurate, comprehensive and regularly updated.”

AI in Instruction Design: L&D’s Personal Assistant

AI can help instructional designers save time in training development, allowing them to deploy and deliver targeted training faster. It also allows them to use the extra time and energy on enhancing the process. However, AI can never replace the expertise of the learning leader.

“… the best uses of AI right now are mostly focused on creating initial drafts, or helping L&D leaders develop processes and systems,” Jacobi shares. “Although exciting, it is worthwhile for leaders to practice discernment when deciding to leverage AI … many of the most fundamental aspects of the training process are still best left to employees who can provide organization-specific insights and a human touch to the design and implementation of training programs.”

“The most important takeaway is that the goal of using AI for L&D is to enhance human capabilities and enrich the learning experience, rather than to replace the human elements that are essential for effective education and training,” Blackmon summarizes.

AI presents invaluable opportunities for streamlining training development and optimizing processes, yet it is essential to recognize its limitations. Although it can expedite initial drafts and aid in system development, it cannot substitute the nuanced expertise and human touch of learning leaders.