The Difference Between Artificial Intelligence And Machine Learning
Artificial Intelligence and Machine Learning are frequently talked about as linchpins for revolutionizing student support and operational efficiency in education. More often than not, however, the two terms are used interchangeably. While Artificial Intelligence and Machine Learning are very similar, there are distinct differences between them. If you’ve ever been uncertain of the difference between the two or what their value in education is, read on.
While the talk around Artificial Intelligence (AI) has certainly ramped up as of recent, the concept has been around for a long time, even since the days of Greek mythology. In a broader context, AI is the concept of machines being able to carry out tasks in a way that humans would consider “smart.” Basically, it’s the science and engineering that make machines behave in ways that initially required human intelligence. A good example of AI is single player chess on the computer. In this case, you’re playing a game of chess on your computer against your computer. The computer has been programmed to make intelligent chess moves so the human player will feel like they are competing against a real person, not a machine. You can even choose the level of difficulty that the computer plays with. For each level, the computer is programmed differently to make it easier or harder to beat. The computer has been programmed with AI to play a chess match and emulate the decisions a human would make.
The concept of Machine Learning (ML) actually falls under the broader umbrella of AI. What distinguishes ML from AI is the “learning” aspect of the term. The idea stems from the notion that humans should be able to provide machines with a set of data and allow the machines to “learn” for themselves. Then, through repeated study of different data sets combined with validation of what is correct and what isn’t (either by sensing if the conclusion is right on their own, or by being told from a third party), the machines will “learn” more about the subject matter through finding common patterns and nuances. In short, ML is designed to learn and develop its own capabilities whereas AI is restricted to the data and instruction human programmers provide the machines.
To dive a bit deeper into how it works, ML utilizes Neural Networks, which are computer systems designed to classify information the same way a human brain does. Through Neural Networks, a machine can be taught to recognize, for example, pictures or images, and classify them according to the features they contain. This system essentially works using probability. Based on data the machine are given, they’re able to make statements, decisions or predictions with a baseline degree of certainty. That degree of certainty is heightened through the addition of a feedback loop, which enables the machine to “learn” by sensing or being told whether its decisions are right or wrong. With this increased certainty, it modifies the approach it takes in the future.
The example I like to give when explaining the difference between AI and ML goes back to the example of computer games. Imagine you’re playing a virtual 5 on 5 basketball game against the computer. Much like in a real game, where the coach of a team can assess the tendencies of the opponent and use that knowledge to make adjustments to counter the opponent’s play, a computer can do the same thing. Throughout the entire game, the computer is gathering data on your playing style, and it changes its behavior to increase its chances of defeating you. If it notices that you use a certain player on your team to score the majority of your points, the computer may elect to double team that player, forcing you to score using someone else, which may be more challenging for you. If this was a best of 3 games series, the computer may start the next game with the double team strategy if it was effective during the first game. These adjustments in response to gathering data signify the “learning” aspect of ML. AI, using the same example, would be the computer being programmed to play in a way that mimics human decisions but won’t “learn” from the data gathered during the current game to impact its strategy for the next game you play against it. won’t change its strategy during the game depending on how you, the user, is playing.
The Education Sector
Now, as far as education goes, what do these two technologies bring to the table? The answer can be its own blog post, or maybe even a series, but here are a few quick examples.
For AI, engineers working with institutions can program a texting system to respond to student inquiries based on keywords in the question. For example, Georgia State University utilizes AI text messages that mimic the role of a college counselor to quickly answer student questions about resources, policies, concerns, and more. GSU has seen a 21.4% decrease in Summer Melt, a term for when admitted students “melt off” during the summer and fail to re-enroll. ML, however, is where the potential for groundbreaking integration lies. Through ML, a system can, in theory, assess a student’s performance on their assessments, and recommend study materials or different learning habits based on their results and/or the questions the student got wrong. With ML, all students will have access to personalized coaching and tutoring based on their behaviors and assessment performances.
There are a number of ways that AI and ML can impact education moving forward. A big reason is the fact that the internet has become so integral to the operations of many organizations and contains endless amounts of information formatted in a way that’s easy for machines to decipher. Since machines have advantages in speed, accuracy, and lack of bias over humans, they will be positioned to take over tedious administrative tasks, freeing up school staff to focus on other areas like student support, and will potentially introduce groundbreaking processes to increase student success. It’s useful to understand the differences between AI and ML so that as we look at ways to build better systems for students and school operations, we’re making the right choices.
Request a Demo