AI and ML: They’re Not Interchangeable

Jan 21, 2020 9:30:00 AM

blog- AI_ML

Those random toilet paper recommendations? That's because of machine learning (not AI).

The “royal we” have taken to calling everything about how machines get smart artificial intelligence. It’s like saying ‘get me a Kleenex’ or ‘can you Xerox this?’ But, in reality, that’s being lazy. There are so many nuances and so many attributes to both artificial intelligence and its litany of subsets - including machine learning - that we have to separate them. The most critical reason is that, well, artificial intelligence does not actually exist yet.

Even though we’re beginning our third decade of the 21st century (I thought we’d all be stowing our flying cars right in our briefcases Jetsons-style by now, tbh), we have a way to go before we have computers that can accurately mimic human thought process and decision-making skills. However, we have made significant strides in making our machines learn from the data and patterns we expose them to and then apply that knowledge to improve things like customer experience. 

Machine learning pioneer Arthur Lee Samuel defines it as “the field of study that gives computers the ability to learn without being explicitly programmed”. (Artificial intelligence is the broader field of study above that.) For example, when Amazon recommends you consider its private label toilet paper, that’s probably because you’ve already purchased its private label paper towels and some other brand of toilet paper and it has learned that, 1) you indeed need toilet paper and, 2) you are open to its private label brand. That right there is machine learning at work. When Netflix recommends you watch The Politician tonight, it’s probably because it knows you liked House of Cards (Robin Wright is so hard-core) and Designated Survivor (although Keifer Sutherland will always be Jack Bauer to me). Same thing. 

The key differences are:

Artificial Intelligence Machine Learning
Broad field encompassing the ideas of creating intelligent machines Sub-field of AI encompassing how machines acquire knowledge and skills by doing
Focused on success Focused on accuracy
Aims to simulate natural human intelligence and decision making Learns from data

 

As we get deeper into smarter computing, and more of our lives become automated - which is both convenient and terrifying - we have to be mindful about what’s what. Because while our computers can arguably look like humans and talk like humans, they cannot yet THINK like humans. When that day comes, we will need the words.

Candace Sheitelman

Written by Candace Sheitelman

Candace Sheitelman brings more than two decades of marketing expertise to Edify, much of it focused on CX and the contact center. She's responsible for Edify's go-to-market strategy and execution. Sheitelman previously ran her own marketing communications firm and global marketing at Aspect. She earned her B.S. in Public Relations from the S.I. Newhouse School of Public Communications at Syracuse University.