What began as destruction of property grew into a direct confrontation with the British Army, and ultimately assassination, execution and transportation.
Now workers are being threatened again, this time by artificial intelligence (AI), with every news site from Forbes to The Guardian screaming that “robots are taking all our jobs”. In truth, robots have been “taking all our jobs” since the 1960s.
Generally, that has not concerned the thinking classes so long as the jobs being replaced were blue collar factory and warehouse automation.
Delivery drones, driverless taxis, and automated lorry trains sit in the same class: threatening those people over there, therefore not a concern to us!
So the latest flap over artificial intelligence as the ultimate jobs killer, is largely a result of an existential panic by the thinking classes. If AI can do the thinking, what are we going to do?While this isn’t exactly keeping the average L&D professional awake at night just yet, you do hear it mentioned during coffee break conversations.
No need to panic just yet. Every innovation since the wheel has had a destructive influence on what went before (the wheel put log rollers out of business), while creating a new wave of creativity and new things for people to do.
Only months into embracing AI and Chatbots at Learning Pool our CTO has suggested we need to hire new employees called Conversation Consultants. And on it goes.
We’ve looked at the Conversation Consultant’s job role, and to be honest, it is a pretty cool new job for Learning and Development. You take your client’s source content and train an AI application to understand it such that the learner can interact with it conversationally.
Ultimately our goal is that learners will talk to their training. More of that later.
The core technology behind this is Natural Language Search (NLS), which has been made widely accessible through AI cloud services like IBM’s Watson. Search used to be about keywords; now it’s about meaning.
Once the quality of search results was directly related to your ability to predict how the data was structured, and how you should ask for the information you needed; now it is about how well the search engine understands your intent and the content it stores.
A simple search for “stock” might give search results for securities, recipes, standards, and punishment devices. You could improve results with additional keywords, excluded keywords, multiple words in quotes, ANDs and ORs. Provided you knew your stem searches from your wild cards, it worked.
We knew no better. Until Google. In 1998 Google’s founding mission statement, “We will organize the world’s information, and make it universally accessible and useful”, started an arms race that has delivered astonishing innovation.
Try this example. Type into Google “I want a flat white near me”, and you will be recommended a coffee shop nearby.
Now search “I want a white flat near me,” and you’ll get property websites. The search engine understands what you mean.
We take a search like this for granted but what’s going on is a really cool process called entity analysis, parsing the text using the same rules a person would deploy subconsciously in their everyday conversations.
Let’s strip that down. The first layer of understanding is keywords. We listen for them in sentences and even if we do not catch the full sentence they give us partial meaning.
You might then look for entities, classes into which words fall. Learning Pool is a company, John is a name, John Smith is a person. A layer deeper and you find concepts, John Smith works for Learning Pool .
Drill deeper again and you can add layers of context, like sentiment. John likes working for Learning Pool .
Once the AI engine has this basic understanding of the content it can interpret questions and attempt to give back the right answer.
This is the machine learning (or AI) bit. Put simply users interrogate content and a set of rules (or algorithm) is applied to deliver the right answer.
The rules are probability based, and an AI engine will adjust those probabilities based on more people choosing one answer over another, or users up-voting or down-voting answers.
It is the wisdom of crowds, the more people choose one answer over another the more the system weights that answer as probably right. Typically, you would seed a data set by indicating probable answers and then let it run.
This is incredibly complex and expensive technology, but services like IBM Watson have made it accessible to the smallest of companies. And tools like Watson Knowledge Studio put the skills in the hands of editors rather than technologists.
At Learning Pool we are using AI to build Performance Support knowledge bases for customer service and outbound sales applications.
In these environments, it is often not possible to excuse people from their desks to do training, but with a Chatbot and an AI-driven knowledgebase, you can train at the desktop, while keeping people productive.
It’s not just about performance support either. Thanks to the structure of our Adapt content we can ingest full training courses into the AI engine.
We asked the eccentric question, “if you could speak to your LMS what would you say to it?” and created an LMS vocabulary that you interact with via the Chatbot.
“What courses am I enrolled on?”, “Do you have any business skills courses?”, “What mandatory learning do I need to complete?”. The bot responds with snippets of information and clickable links which launch full courses.
This is the mythical invisible LMS made real, and it is very exciting to see it working.
One thing we’re certain of is that it is not going to eradicate any learning and development jobs; it will create further challenges, opportunities, and the need for people to solve them.
So, no need for smashing of looms, assassination, execution and transportation this time; just retraining, upskilling, and more creative thinking.
If you’re interested in finding out more on this subject or any of our products please fill out the form below and we’ll be in touch.
Paul Healy has worked in the learning industry since 2003 in sales, learning consultancy, and programme management. He specialises in assisting companies with change management and innovation agendas.
How many employees require elearning?
+44 (0) 207 101 9383
+44 (0) 207 101 9383
US (857) 284-1420
+44 (0) 845 074 4114*
* call charges vary depending on your provider