Many of the products and services we use every day are already leveraging Artificial Intelligence (AI) to improve the user’s experience.
Amazon uses machine learning to recommend products to you based on information it has collected. Google Home, Apple’s Siri and Microsoft’s Alexa use AI heavily in speech-to-text conversion and optimization. Your email service probably uses AI to fight spam emails and prevent them from landing in your inbox.
The use of AI is all around us and can bring great benefits to the Learning and Development (L&D) sector and ultimately, the workforce also.
L&D professionals need to stay on top of fast changing technology to optimize the learning experience and outcomes, developing new learning strategies and methodologies that take advantage of these improvements, especially when it comes to AI.
For example, a Gartner report predicts that AI bots will be powering 85% of customer service interactions by 2020 and another report states that 20% of business content (including training content) will by this time be written by AI. In addition, the Bank of America predicts that AI will drive between $14-33 trillion annually of economic growth by 2025.
AI is going to have a huge impact on the L&D industry. Organizations have a huge amount of data available to them, which they can analyze and use to optimize training programs and learning curriculums.
Gone are the days when every employee needs to learn the same course content. Content can be personalized to suit the learner’s needs, focus on weaker areas of the learner, recommend suitable content based on past behavior, predict needs based on their role, and even auto-generate content using various content creation algorithms.
In order for AI to be utilized fully, organizations need to harness the huge amounts of data using machine learning, data analysts, AI programmers, and more. The output from this data enables L&D departments to gain insights into the learner journey and helps them to create training programs that drive value and enable adaptive learning.
Learning styles impact the development of learning solutions. A person’s learning style may be influenced by age, ethnicity, cultural background, and other factors which must be considered in the development process.
For example, a study by the University of Georgia demonstrated that "males scored significantly higher on the Abstract Sequential channel than females, and women scored significantly higher on the Abstract Random channel than males", showing that a different teaching style (which could be optimized by AI) would benefit each gender.
Millennials now rate training and personal development as the number one job benefit. Organizations need to recognize this and start deploying AI to train employees by optimizing content to suit the users preferred learning style. This will not only make the learning experience more enjoyable for them, but help with knowledge retention and on the job performance.
As discussed, each employee will have a preferred learning style and learn most effectively using a specific method. This could be through video tutorials, written content, in-person training, gamification, audio guided presentations, or something else.
An AI powered training program allows the training program to be adaptive, where the modules are modified to suit the needs of each employee. The LMS might offer video tutorials to certain employees, but auto transcribe the videos to text-based articles for other employees. It could be able to create visuals based on written content and suggest the employee take an in-person training day on sections of the course they are struggling with.
Learning insights also help develop a wider understanding of learner behavior, leading to predictive capacities. Using the insights, organizations can create intelligent and smarter positioned content, that’s adaptive, intuitive, and responsive to a learner’s personal journey.
Tests, quizzes and assessments are becoming an important part of e-learning. They help consolidate learning and measure the effectiveness of the learning course. However, one limitation of these assessments is they take a one-size-fits-all approach.
The tests are developed along the line-of-best-fit to ensure that the tests are suitable for all learners whatever their capability. This means that some learners may find such tests too easy and others may find them too difficult.
AI allows us to design adaptive assessments that go beyond the static Q&A format. AI has the ability to assess individual ability and progression, tailoring the subsequent course content based on the results of these assessments.
Iris powers the PluralSight assessment algorithms and guides users to the skills they need.
An example of this is Iris, developed by the technical training provider PluralSight: "Iris... updates certainty, question difficulty and skill ratings as she collects feedback. Using natural language processing and machine learning, Iris recommends content based on your Skill IQ."
AI can dramatically shorten the learning process by suggesting only specific modules the employee needs to improve the skills for the job they are doing. The system, having learnt the user’s strengths, weaknesses and learning preferences, can suggest suitable training courses and modules for the employee.
An example of this is the personalized course recommendations from LinkedIn Learning.
AI can improve the employee training experience and provide feedback on areas to improve.
For example, using tools such as speech-to-text which rely heavily on AI and ML, users can receive feedback on their presentation performance in areas such as pace of voice, number of hesitation words used, whether certain keywords were mentioned, and so on.
VirtualSpeech is one of the companies taking advantage of AI, combining it with VR to give users a realistic way to practice different soft skills and provide instant feedback on their performance.
Completion rates for online training are poor. Data for MOOCs shows that less than 15% of people actually complete the course.
With machine learning and trained AI, the system can provide only relevant training resources and content in the format the learner wants (e.g. video based). This should dramatically increase completion rates for the training courses and ensure better learning outcomes for employees.
AI will have a huge impact on accessibility for people with disabilities. It takes many additional hours to develop training content which can be accessed by everyone, so able-bodied people tend to benefit the most from training.
AI opens up training to more and more learners, irrespective of any disabilities they may have by enabling functioning senses the learner has.
For example Microsoft’s Seeing AI app gives us an idea of what is to come. The app "narrates the world around you. Designed for the low vision community, this research project harnesses the power of AI to describe people, text and objects" - in effect, it allows the partially sighted to use their phone as a third eye.
With this kind of AI augmentation, learning developers will eventually be able to design courses that are more broadly accessible.
We’ve seen rapid progress in this field already, with Google introducing Automatic captions in YouTube - "multiple caption tracks, improved search functionality and even automatic translation", which "not only help the deaf and hearing impaired, but with machine translation, they also enable people around the world to access video content in any of 51 languages". This functionality makes YouTube much more accessible to millions of people around the world.
Performance data can be collected when employees are performing their jobs in the workplace and also during training courses. Once this data is collected, it can be analyzed and AI applied to gain certain insights.
When workplace data is analyzed, L&D professionals gain insights into which training programs should be assigned to specific employees to improve their effectiveness at their job. Data analysis should highlight areas employees need to improve and the AI system can then recommend suitable courses to fill the knowledge gap.
Analyzing training data reveals which types of training material should be assigned to employees based on their preferred learning styles. It can also help benchmark employees against each other with assessments and provide a suitable difficulty of training material to challenge the learner.
Any gaps uncovered in the training can be assessed and training programs re-designed.
Collecting and analyzing employee data needs to be continuous, so that systems stay up to date with current trends and AI models trained with the latest data.
AI-based tutors can improve learning efficiencies, even compared to experience in the field. Several years ago, DARPA sponsored "the development of a digital tutor that uses AI to model the interaction between an expert and a novice" to try to reduce the time taken for Navy recruits to become experts in technical skills.
The study concluded that Navy recruits using the digital tutor to become IT systems administrators frequently outperform Navy experts with 7-10 years of experience in both written tests of knowledge and real-world problem solving.
The study also suggested that "workers who have completed a training program that uses the digital tutor are more likely to get a high-tech job", and that "research that enables the emergence of an industry that uses AI approaches such as digital tutors could potentially help workers acquire in demand skills."
Researching, finding and building training programs is a time consuming part of the L&D process. Often organizations resort to buying in training content to save time.
For example, a learner might want to understand how a particular process on a turbine engine works - the AI system then crawls the web for research papers on the topic and provides relevant snippets from these papers to help the learner understand the process.
AI can also locate and extract appropriate information and convert it into human readable formats suitable for learning. A basic version of this can be seen on YouTube in the form of the computer-generated news stories with video, pictures and spoken script which are put together by AI tools such as Wibbitz, using data from across the web.
This automation means that AI can do a lot of the heavy-lifting of the course design process, leaving instructional designers with more time to focus on delivery and design.
Whether we realize it or not, we may be prejudiced against certain parts of the L&D process. This can range from creating a new course with only a few assessments because we don’t particularly like them, to grading students differently based on their background, to accepting new hires based on how we felt about them.
AI enables the L&D process to become impartial and objective, with processes, learning paths, new hires, new courses, etc. based on data and results.
However, it’s important to be aware that bias can still creep in when training the AI and ML models, as predefined variables and datasets are still chosen by humans.
Just like how if a company hires people from all different backgrounds, new ideas will be varied and different, if a company’s AI team is diverse, the chances of bias will be reduced.
To remain both accurate and relevant, the L&D AI system needs to be continually trained to account for changes in the market and remove any bias that occurs through new data sets.
Read more about bias in AI:
In our fast-changing world, L&D and HR teams have to be proactive. They need to ensure the most relevant training tools and knowledge resources are available to their employees when they want to learn.
The technology landscape is changing extremely rapidly. Organizations need to stay on top of the latest trends, whether that be VR or Blockchain, in order to stay relevant. By upskilling workers within your organization, you’re able to keep evolving with changing technology.
A key part of this upskilling is through the use of AI and ML to provide employees with the most relevant content when they need it. AI will transform how learning content is delivered, leading to greater alignment with business values. This requires L&D leaders to understand the latest AI and ML methodologies and trends.