Criticism can be painful. But criticism - and how we respond to it, using emotional intelligence - is one of the richest kinds of fuel for personal growth.
How to explain machine learning in plain English
What is machine learning? What is ML vs. AI? What data problems could ML solve for your organization? Here’s how to discuss the key issues in plain terms
What machine learning can do: How to evangelize
Some explanation is likely necessary, but the longer-term task is better described as evangelism: Extolling the benefits of ML rather than getting into the technical nitty-gritty, as Fernandez advises above.
For many companies, a starting point in identifying those benefits – or the problems ML can solve – again comes down to data. This time, though, it’s a matter of the massive amounts of data so many organizations are generating. It’s getting harder and harder for humans alone to make sense of it on their own, at least in an efficient manner.
“The ability of machine learning models to deal with high-dimensional data is extremely helpful for businesses,” Brock says. “AI-enhanced software can do things such as finding patterns in user access data and accurately predicting customer retention, which would otherwise be impossible for a human to do.”
McCourt takes a similar tack when it comes to explaining the potential importance and value of data in the business world. It’s about making sense of data that might otherwise be indecipherable in a timely manner.
“In a business context, I usually talk about ML as a tool that can provide insights into complicated (or simple) circumstances,” McCourt says. “ML gives a mechanism to make smarter decisions by more completely and effectively using the data which is available to your company. It also gives a mechanism by which certain choices can be automated, freeing up resources for teams to make more complex decisions.”
Machine learning’s superpowers: Learning from practice, detecting patterns
When trying to help non-technical audiences understand the basic concepts of machine learning, McCourt likes to note that machines learn much like humans do: with practice.
“Machines need a structure in which to exist (a set of rules) a goal to achieve (to measure success and provide feedback), and opportunities to fail and learn (iterating over data),” McCourt says.
Finally, when evangelizing ML to a broad audience, McCourt finds success in emphasizing that machines have the potential to identify patterns that humans would not otherwise see. It’s not about machines replacing humans per se, but about creating new possibilities that didn’t exist previously.
“Machine learning is often applied with the goal of identifying new behavior and taking smarter action. At times, this can mean that tedious tasks on which a human would normally be wasted, such as flagging images or posts with offensive content, can now be delegated to a machine,” McCourt says.
In the past, writing rules (i.e., writing code) to teach a computer the parameters of what’s offensive and what’s OK would have been challenging at best, McCourt notes.
“Now computers can be shown a stream of content from which they learn how to make appropriate distinctions,” McCourt says. Longer-term, that’s hopefully a relatively mundane use case too. “Ideally,” he adds, “this goes even further than this example, allowing humans to leverage ML to solve problems that humans solve poorly or at great cost, such as diagnosing illnesses or predicting equipment failures.”
[ What AI can and can’t do now: AI in the enterprise: 8 myths, debunked. ]