Science fiction literature is rife with musings on the impact of Artificial Intelligence (AI) on human life. Perhaps the best portrayal of how that would play out in the real world was the movie A.I. Artificial Intelligence (2001), directed by Steven Spielberg.
Constructing AI is probably the most ambitious human undertaking ever conceived. Simply put, AI is sentient intelligence that could simulate human behavior and think to such an extent that you would intuitively recognize it as an autonomously thinking entity, as you do with other people on a daily basis.
However, do notice that many assumptions are embedded in the AI concept:
- The assumption that we understand what is sentience.
- The assumption that we understand how the human mind works.
- The assumption that we understand why humans behave a certain way.
- The assumption that sentience presupposes some form of free will.
It doesn’t take long to grasp that we still lack a proper understanding of the human mind, and the terms we use to describe our mind’s operations. And if we don’t understand that, how can we possibly hope to ever develop a fully-featured AI from the sci-fi movies?
For the time being, you might get fooled by a particularly clever chatbot, or bested by a super-computer in a video game, but there is no evidence to suggest there ever will be AI as we have come to understand it from science-fiction.
Grounding in Reality
Instead, what is far more interesting for our immediate reality is machine learning. On the path of researching how to recreate human thinking, many successful machine learning algorithms have been developed. Machine learning simply means that flexible algorithms can be deployed to solve certain problems without being explicitly programmed to. As they are fed data, these algorithms can detect patterns and accumulatively learn.
Machine learning algorithms are already in wide use with almost everything concerning language, such as:
- Matching your interests with ads;
- Smart spam and malware detector;
- Text prediction;
- News generation;
- Contextual translation;
- Customer support.
In non-linguistic applications, the usage is similarly wide-ranging:
- Traffic detection, prediction, and navigation;
- Agricultural crops and livestock management;
- Video surveillance and facial recognition;
- Gait recognition.
Yes, that last one means that AI bots can discern an individual even based on how he/she walks. It stands to reason, as we humans can do it as well, but with much lesser consistency. As you can see, we are immersed in a world of data, and if that data can be quantified by programming engineers, a machine learning algorithm can as well.
The question is, can machine learning improve our children’s ability to learn? Certainly, any essay writer would benefit from an objective machine mind, the same as any student.
Impact of Machine Learning on Education
Given the fact that the entire edifice of education is based on linguistics and symbols, a natural environment for machine learning algorithms, it stands to reason that we can expect to see a massive impact of machine learning in all educational institutions. More importantly, within homeschooling networks and online platforms.
Let’s face it; it is exceedingly rare to find a teacher that can truly be flexible enough to adapt to every student’s learning needs. It’s simply the way we humans are constituted. Once we have a way of doing things, we rarely veer off that path. With machine learning, there is no such obstacle in the form of calcified habits.
Unbiased, Cheap and Efficient Education
Based on the continuous feedback the student gives to the program, machine learning can adaptively customize learning lessons and methods. Called adaptive learning, this machine learning algorithm constantly assesses the performance of the student and modifies its teaching algorithms. No mutual student-teacher frustration, no ill-will, no embarrassment – just pure efficiency until the lesson is adequately completed.
More importantly, this means that every student can have a personal tutor, but without the cost of one. Keep in mind, only upper-middle-class and rich families could afford personal tutors for every subject. And even then, it often is not enough due to human communication problems and personalities.
Unbiased exam assessments would be another huge boom to education. We already know from multiple studies that female teachers rate boys poorly compared to girls. Female teachers had a tremendous negative effect on boys and their employment prospects, which is another negative that could be completely removed by machine learning.
Needless to say, machine learning advances will likely force millions of people to change professions or adapt in other ways, but how can anyone stop technology with such obvious net positives? That line of thinking certainly didn’t stop automobiles from overtaking horse carriages.