Editorial Note: This article was updated in March 2022 to reflect the latest AI and workforce trends.
Back when this article was originally written, 2020 was the year that cars would fly, time travel would be possible, and life could be extended indefinitely. Endeavors that may have started off as purely science fiction were starting to bear fruit. As the stampede scientific development marches on, one technology that has not lived up to the hype is Artificial Intelligence (AI).
While true AI does not yet exist, machines can learn, which we explore in this article on 'How computers learn'. Let's start by breaking down what artificial intelligence is.
What actually is Artificial Intelligence?
AI means a lot of different things to a lot of different people, so let's break it down. According to Merriam-Webster, it is:
- a branch of computer science dealing with the simulation of intelligent behavior in computers;
- the capability of a machine to imitate intelligent human behavior.
Put into practical application, AI is the ability for a machine to react to a situation it has never experienced or learned about
Think about dropping a Los Angeles alley cat into the Kruger National park in front of an elephant. How would the cat react? It has never seen an elephant in LA, nor been on one of the plains of Africa. Now compare this to dropping a human into a similar situation: the human could generalize that because they once saw an elephant in a zoo, the human knows that the elephant is not interested in eating them.
Generalizing is not yet possible for a computer. Computers ‘learn’ by ingesting massive amounts of data about a specific situation. Algorithms that allow computers to generalize across unrelated topics are just starting to emerge and are the first step in a long process that may eventually end in computers being self-aware. What has gained prominence in the workplace is machine learning (ML).
So what is Machine Learning?
The Brookings Institute does an excellent job of delineating ML from AI: “The core insight of machine learning is that much of what we recognize as intelligence hinges on probability rather than reason or logic.” In application, ML is the use of statistical, actuarial, and other mathematical models to identify trends at scale in large datasets.
AI, ML, and your job
It is impossible to know how the workforce will change if computers become rational-thinking beings. For now, rest assured fellow humans, ML is not coming for your job. There are three main reasons for this:
- ML relies heavily on manual (code for human) data gathering and preparation in order to ‘learn’;
- Computers require humans to build the algorithms that power ML;
- ML allows computers to react to well-defined problem sets (such as immediately determining if a credit card transaction is fraudulent) but requires humans for quality control and to resolve the outcome of the inference.
Machine Learning in the workplace
Imagine a world where you have the ability to predict an events’ likely outcome. This ability exists and is already being applied equally across the financial, retail, manufacturing, logistics, and agricultural sectors for well-understood subject areas. Let’s look at one of the practical and proven uses of ML, driving sales. Here is the situation: an insurance agent answers a client's question. Seems simple right? In an article on Sales and Insurance, we dive into this topic even more.
During a routine interaction, the insurance agent casually the client ten potential coverages which they do not already have. The individual is only interested however in one of the nine offered policies. As a result, the client comes away from the encounter with a feeling of being overwhelmed, and 'sold to.' For any insurance professional, this is the least desired outcome as it can damage the trust between the two parties. And in most situations where a client feels sold to, their instinct is just to say “NO” to everything - even the policy they would have otherwise been interested in. What a lost opportunity!
With machine learning insights, the conversation would have gone differently. In the near future, insurance agents will regularly use ML to provide context about their client's buying behavior. If they now can know that their client has a 92% chance of needing a commercial property policy, does the agent still waste time offering the client the other nine policies that there is zero chance they’ll purchase? Of course not. Within this context, the ML informed the agent that their client would need additional coverage. The agent only presents additional lines that would benefit the client who can then easily review these coverages. Both parties win.
The statistical analysis at scale made possible by ML will enable companies to serve their clients and provide valuable insight during interactions with customers.
The statistical analysis at scale made possible by ML will enable companies to serve their clients and provide valuable insight during interactions with customers. We think of this as an 'AI Buddy System,' which we explain in this article. So while an individual’s job may see some change as the result of ML insights, machine learning will not be replacing them.
Future of Work à la ML
Technology is leading to changes in how the workforce spends eight hours a day for their career spanning 30+ years. Naturally, this change has led to uncertainty about the role of technology, specifically machine learning (or AI), in replacing workers. Understanding the capabilities of ML is a starting point to understanding how this technology is enabling workers, not causing job loss. A couple ways ML is changing the way businesses operate:
- Increasing retention by targeting clients with customized loyalty initiatives.
- Personalizing customer service to maximize every client encounter.
- Identify fraud early in the underwriting process, reducing costly claims.
- Smoother supply chains for just-in-time logistics.
Far more likely to contribute to systemic job-loss is new consumer preferences, technological advancements (like using nuclear fusion for power generation rather than coal), and the practice of offshoring jobs.
I've had the pleasure of speaking with many seasoned insurance professionals resistant to our ML because they feel threatened; they think this advancement will lead to job losses. This could not be further from reality. Ultimately, the promise of machine learning is efficiency. The average workday will contain far less mundane tasks, and create more enriching work experiences.
Because no one enjoys crunching numbers in an Excel sheet - wouldn’t you rather engage in meaningful conversations with customers you have a greater than 51% chance of making the sale to?
