Helping learning and development climb the artificial intelligence ladder

A little over a year ago, I stood in the expo hall at a learning and development (L&D) conference with JD Dillon, founder and principal at LearnGeek. We watched demos of chatbots, predictive learning management systems, and gamified courses. As vendors touted their artificial intelligence (AI) solutions as “AI-enabled” to power “the future of work,” we joked about the appropriation of AI as a sales tag and laughed at the perpetuated myths about what AI can and can’t do today.

Still, there was no denying that artificial intelligence was becoming the next significant conversation in workplace learning. Given the impact of AI on business, it’s no surprise. The IDC predicts that 75 percent of enterprise applications will use AI by 2021 and forecasts that cognitive and AI spending will grow to $52.2 billion. Recently, JD and I co-presented, “Helping L&D Climb the AI Ladder” with a goal of advancing the AI conversation within L&D. We believe that it is time that L&D collectively explores the range of AI capabilities and the practical steps it can take to improve its practices and, in doing so, climb the AI ladder.

What is AI?

Whenever a machine completes tasks based on a set of stipulated rules that solve problems (algorithms), this “intelligent” behavior is called AI. Most experts differentiate AI into narrow and general as a way to explain where we are today versus where we’ll be in the future.

General AI is a revolutionary state when AI thinks and functions like humans, capable of perceptual tasks like vision and language processing as well as cognitive tasks like thinking and understanding. The tools and technology to create general AI don’t exist today and it’s years away.

Narrow AI, however, is emerging. We use this type of AI regularly: Spam filters, recommendation algorithms, image tagging, or voice assistants like Amazon Alexa. Narrow AI can execute highly focused tasks: pattern recognition, natural language processing, conversational response, discovery, visual recognition, sentiment analysis, and text to speech. These applications use machine learning, a method of data analysis, to train algorithms using data.

With these capabilities in mind, consider the question, “Which is a better AI application: smart traffic lights or self-driving cars?”

The answer depends on whether you’re trying to improve the existing traffic light system or reimagine transportation. Traffic lights are an intervention to manage the decision-making process of individual drivers. Making them smart could increase the efficiency of traffic flow, reduce emissions, and decrease accidents.

Self-driving cars, on the other hand, exist in a world where drivers don’t make decisions, but where the responsibility for efficient traffic flow is distributed across a network of nodes that respond to real-time data. In the world of driverless cars, why focus on optimizing traffic lights?

Applying AI to L&D

The L&D industry has historically responded to emerging technology by focusing on the traffic lights, taking the stuff it has always done and re-publishing it in a new format. Classroom training became webinars and eLearning; course catalogs became learning management systems; training manuals became PDFs and ebooks.

We can no longer perpetuate this cycle. AI represents the next potential inflection point for L&D and an opportunity to truly change learning in the modern workplace. It is time for L&D to focus on the world of driverless cars and harness the potential of AI to predict, enable, reimagine, and automate workplace learning.

Imagine an application that can author content based on product documentation and discussion threads. Imagine that content being translated in real-time, then uploaded, tagged, and made accessible to everyone in an organization. Recommendation, administration, translation, natural language processing (NLP) search, and content tagging are being implemented right now. Using AI for gap identification, impact analysis, personalization, mentoring, coaching, and chatbots are the next wave.

But the bottleneck for L&D is that AI is not just plug-and-play. These applications are powered by rich data, and that doesn’t come from a box. If L&D wants to realize the full potential of AI, then we have to fix our data practices.

There is no AI without Information Architecture

In Forrester’s Predictions 2019: Artificial Intelligence report, 60 percent of the decision makers at firms adopting AI cite data quality as the number one challenge when trying to deliver AI capabilities. In most organizations, data is all over the place, spread across different platforms in different formats with no interoperability. That’s why the first rung of the AI ladder is information architecture (IA): the collection, organization, structure and labeling of data. IA enables analytics, which enables machine learning, which powers AI. In other words, “there’s no AI without IA.”

To start climbing the AI ladder, L&D must improve the quantity and quality of learning and performance data that we collect to power high-value AI: demographics, consumption, feedback, context, connections, knowledge, behavior and results. This leads to the second rung of the AI ladder: plug into the broader organization.

Since L&D is not likely introducing AI to the workplace, leaders need to figure out who is, get in on the ground floor, and partner on the new workflows and processes necessary to collect, clean, store, analyze, and access that data.

Artificial intelligence is changing the way work gets done, including the practices that L&D uses to help people do their best work every day.

AWS Training and Certification helps individuals and employers build and validate the cloud skills required to transform business with AI. As players in the modern learning ecosystem, we are all responsible for imagining how new capabilities can reinvent the world around us and for investing in the infrastructure that enables the future of work.