Can Machine Learning Personalize Education Without Isolating the Learner?

“Personalized learning” seems to be having an identity crisis of sorts. One group defines it one way, while another group of educators, schools, and technology developers uses the term to describe something different.

That lack of agreement appears to be one of the main problems surrounding personalized learning. We do not have a communal definition on which to base a shared set of standards. If we cannot agree on what personalized learning is, how can we expect machine learning to advance the experience we are trying to create?

It begins to sound like a comedy sketch built around a failure to communicate:

“Who’s the first baseman?” asks Costello.

“Who is the first baseman,” Abbott replies.

“That’s what I asked. Who is the first baseman?”

“Who’s on first base!” Abbott exclaims.

“I don’t know. What are you asking me for?” Costello shouts.

It would be quite farcical if it were not also true.

Before discussing how machine learning can make a new form of personalized learning possible, we first need to define what we mean by personalized learning.

  • Do we mean individualized or customized learning paths?
  • Do we mean using technology to support instruction?
  • Do we mean one-on-one learning?
  • Do we mean student-led experiences in which learners choose the topics they are interested in?

Researchers have pointed out that personalized learning varies widely in both definition and implementation. Teachers, schools, districts, states, and technology developers may all use the term, yet they may be describing very different practices. That variability makes it difficult to study personalized learning consistently or determine which approaches reliably improve student outcomes.

This matters because machine learning can only support the goals we give it. If the educational purpose remains unclear, the technology may simply automate the confusion.

Machine learning can certainly support personalized learning. It can analyze patterns in student performance, identify areas of difficulty, recommend resources, adjust the sequence or difficulty of content, and provide feedback based on a learner’s progress. It may also help instructors recognize when a student is struggling before the problem becomes obvious.

The technology, however, is only as useful as the educational goal guiding it.

Instead of tossing money at educational projects because they use the right buzzwords or promise to “revolutionize the school district,” we should first determine what we want to accomplish and then decide whether machine learning is the appropriate way to support that goal.

The question should not be, “How do we use machine learning because it is available?”

The better question is, “What learning problem are we trying to solve, and can machine learning help us solve it responsibly?”

Should Personalized Learning Be Applied to Every Subject?

No.

Life is a communal gathering, and school teaches us more than academic content. It teaches us how to live in society, communicate with others, accept different beliefs, benefit from other people’s input, and listen to opinions with which we may not agree.

If we create the expectation that students will always receive information specifically tailored to their preferences, how will they respond when they encounter obstacles or when their cupcake frosting is vanilla instead of chocolate?

For certain subjects and situations, personalization makes sense. A student who needs additional support in mathematics may benefit from an adaptive system that identifies specific gaps and provides targeted practice. Specialized programs connected to high-demand careers should be designed to resemble the actual career field as closely as possible. Students may also benefit from options that allow them to pursue interests, work at an appropriate pace, or receive resources in formats that better support their learning.

Other subjects and experiences should remain communal.

Students need the traditional classroom in which one child is chewing on a pencil, another is pretending to read while turning the pages backward, and the teacher is doing her best not to have an in-class breakdown.

That atmosphere is a microcosm of what awaits students once they enter the real world.

They will work with people who think differently, move at different speeds, communicate poorly, interrupt, misunderstand directions, or require patience. Not every project, workplace, relationship, or responsibility will be tailored to their individual preferences.

So, let us prepare students for that reality rather than convince them that life will always be customized around their needs and wants.

Finding the Right Balance

Personalized learning should not mean isolating students inside individual digital pathways.

It should mean using available information to provide the right support while preserving opportunities for collaboration, challenge, discussion, and shared experience.

Machine learning could help create that balance. It might quietly adjust practice activities, identify misconceptions, recommend additional support, or give teachers a clearer view of student progress. The student could still participate in a common course, work with peers, and encounter ideas that were not selected solely because they matched an existing interest.

The teacher should remain central to this process. An algorithm may identify a pattern, but it cannot fully understand the family pressures, emotional concerns, cultural experiences, motivations, or relationships affecting the learner. Machine learning should inform instructional judgment, not replace it.

There is also a danger in allowing the system to become too prescriptive. If a platform constantly recommends what a student is most likely to succeed at, it may unintentionally narrow that student’s opportunities. A learner should not be prevented from attempting something difficult simply because an algorithm predicts a lower probability of success.

Personalization should expand options, not quietly restrict them.

Final Thought

Machine learning makes personalized learning possible by allowing educational systems to respond to patterns in student data at a scale that would be difficult for one instructor to manage alone.

But possibility does not equal purpose.

Before transforming educational culture, we need a clearer definition of personalized learning and a shared understanding of when it is appropriate. Some learning experiences should be individualized. Others should require students to participate in a community, encounter frustration, consider unfamiliar ideas, and learn alongside people who are different from them.

The goal should not be to create an educational experience in which every condition is perfectly tailored to the individual.

The goal should be to provide enough personalization to support the learner without removing the shared challenges that help prepare that learner for life.

How Can Learning Analytics Better Support Community College Students?

What learning analytics would be most useful in a community college, and how could those data help students make more informed decisions throughout their educational journey?

At the institution where I worked, we served as an open-enrollment college that accepted students with a wide range of academic backgrounds and life experiences. Placement measures such as the Texas Success Initiative assessment helped determine where students should begin, but those scores only told part of the story.

A large percentage of the student population was first-generation, nontraditional, or both. Many had never attended college, did not have family members who could guide them through the process, or were balancing school with employment, family responsibilities, and financial pressures.

The challenge was not always that students could not understand the course material. Often, they simply did not know the many small but important details involved in navigating higher education. Registration deadlines, degree requirements, grade point average expectations, financial aid rules, scholarship conditions, syllabus responsibilities, and program prerequisites can become overwhelming when no one has explained how the system works.

For that reason, the most useful learning analytics would be those that help identify risk early while also making the student’s academic pathway more visible.

Early-alert data

Attendance, grades, missing assignments, course participation, and sudden changes in performance could all contribute to an early-alert system.

If a student stops attending class, misses several assignments, or experiences a significant decline in grades, advisors and faculty could intervene before the situation becomes difficult to recover from. The purpose should not be to label the student as unsuccessful. It should be to identify when support may be needed.

These data become especially useful when combined with human judgment. A dashboard may show that a student has missed two classes, but a conversation may reveal transportation problems, work conflicts, childcare issues, confusion about the course, or a misunderstanding of institutional procedures.

Analytics can identify the signal. People still need to understand the cause.

Greater transparency for students

Treca Stark of Prince George’s Community College has noted that community colleges operate with limited financial and human resources while carrying multiple responsibilities related to academic preparation, workforce development, and open access. These constraints make it understandable that institutional leaders would explore how analytics could help staff use their time more effectively.

One of the best uses of analytics would be to provide students with greater transparency throughout their college journey.

In previous roles, I worked with many students who did not understand the requirements of their degree plan, how many credit hours they had completed, when they might graduate, or what prerequisites were required for a program they hoped to enter. This confusion often created frustration for the student and additional work for advisors and support staff.

A well-designed dashboard could make this information visible in one place. It might show:

  • Completed and remaining degree requirements
  • Current grade point average
  • Credits earned and credits still needed
  • Progress toward graduation
  • Financial aid status
  • Registration holds
  • Upcoming deadlines
  • Program prerequisites
  • Recommended next steps

This would not remove the need for advising. It would allow the student and advisor to begin the conversation with the same information.

Course recommendations

Another useful approach would be a recommendation system similar to the model associated with Austin Peay State University. Such a system could examine a student’s academic profile, intended program, completed coursework, and the prior success of students with similar backgrounds.

The system could then recommend courses in which the student is more likely to succeed and that keep the student moving efficiently toward completion.

This could be particularly valuable for students who are unsure which course to take next or who unknowingly register for courses that do not apply to their degree.

However, recommendations should remain advisory rather than deterministic. A student should not be restricted because an algorithm predicts a lower probability of success. The system should inform a decision, not make the decision on the student’s behalf.

Why these analytics matter

The greatest value of learning analytics is not simply that they produce more data. Their value lies in whether they make the educational process easier to understand and support timely action.

For community college students, useful analytics could help answer practical questions:

  • Am I currently on track?
  • What do I need to complete next?
  • Am I at risk of losing financial aid?
  • Which course should I take?
  • How close am I to graduation?
  • Who should I contact when I need help?

When students can see their progress clearly, they are better positioned to take responsibility for it. At the same time, faculty and staff can focus their limited time on students who need direct support.

The burden of responsibility becomes more evenly shared. The institution provides clear, timely information, while the student gains a better opportunity to understand and manage the path ahead.

Final thought

The most useful learning analytics for an open-enrollment community college would combine early-alert indicators, degree-progress information, and carefully designed course recommendations.

These tools would not replace advisors, faculty members, or student support staff. They would strengthen their work by making important information easier to see and act upon.

The central question is not whether a college can collect more data. It is whether that data can help students feel less confused, more informed, and better supported as they move toward completing their education.

What Skills Will Tomorrow’s Workers Need Most?

If you were responsible for preparing tomorrow’s workers for a retraining program, which skills would you prioritize?

The possible answers are extensive: emotional intelligence, creativity, cultural flexibility, technological adaptability, storytelling, design thinking, data analysis, financial literacy, mindfulness, entrepreneurship, and many others. My own five choices come from a mixture of what is needed to become a mature, well-rounded individual and what is required to perform effectively across a wide range of workplace situations.

The five skills I would emphasize are:

  1. Self-awareness and self-assessment
  2. Empathy and active listening
  3. Patience and perseverance
  4. Problem-solving
  5. Data analysis

Self-awareness and self-assessment

The ability to assess yourself honestly means that when you look in the mirror each morning, you tell yourself the truth.

You recognize your strengths and weaknesses, your inhibitions and impulsive moments, and the experiences that helped shape the person you are today. When improvement is needed, and there is always room for improvement, self-awareness allows those truths to become visible so that meaningful changes can be made.

This kind of truth-telling begins at the individual level and then flows outward into how we work with others. Employees who understand themselves are more likely to recognize when they need help, receive feedback without becoming defensive, and take responsibility for their decisions.

Empathy and active listening

The ability to listen attentively to another person is becoming increasingly difficult because of the amount of noise surrounding us.

It is easy to pick up a phone, put in an earbud, and tune other people out. Yet human beings are social creatures. We need others with whom we can exchange ideas and share our thoughts, concerns, fears, and anxieties.

Empathy requires us to consider what another person may be experiencing. That begins with listening closely enough to understand their position rather than simply waiting for our opportunity to respond.

In the workplace, active listening supports collaboration, strengthens trust, reduces misunderstandings, and helps teams recognize concerns before they become larger problems.

Patience and perseverance

“Patience is a virtue” may be an old expression, but it remains just as relevant today, perhaps even more so.

It is easy to pursue a skill, project, or goal and then give up when the work becomes difficult or the effort required feels too great. A workforce that understands the patience needed to persevere through challenges develops a mindset that is steady and less likely to be defeated by obstacles.

Retraining itself requires perseverance. Workers may be asked to learn unfamiliar systems, change established habits, or begin again in areas where they were once confident. The ability to tolerate discomfort and continue moving forward may matter as much as any specific technical skill.

Problem-solving

Problem-solving may seem like an obvious choice, but it deserves to be stated directly.

When you see a problem, do not ignore it.

This skill connects back to self-awareness and self-assessment. Rather than avoiding responsibility or kicking the can down the road for someone else to handle, employees should develop the will, patience, and judgment needed to address the issue.

Effective problem-solving requires more than reacting quickly. It involves defining the problem, gathering information, considering possible causes, evaluating alternatives, and determining whether the solution actually worked.

As technology and workplace demands continue to change, workers will increasingly encounter situations for which no established procedure exists. Their value will depend partly on their ability to reason through uncertainty.

Data analysis

If the workforce is expected to progress, compete, and make responsible decisions in a global market, the ability to interpret data will be essential.

Data alone is not knowledge. It must be examined, placed in context, and translated into useful information. Workers need to recognize patterns, question assumptions, distinguish meaningful evidence from noise, and explain what the findings mean for the organization.

The previous four skills connect directly to this ability.

If you are honest about who you are, you will recognize when you need assistance. If you actively listen to the people whose help you seek, you will benefit from their knowledge. If you remain patient while applying that advice, you will continue to grow. As your ability to solve problems develops, you will become better prepared to analyze data and transform it into a valuable resource: information.

And to think, the entire process began with looking in the mirror.

How I Would Implement the Training Program

The training program should combine subject-matter expertise with a deliberate instructional design process.

One approach would be to bring in a freelancer or external specialist to serve as the subject-matter expert. That expert could work with the hiring team, organizational leaders, and current employees to identify the skills and performance gaps that need to be addressed.

From there, the organization could follow an instructional design process:

  • Analyze workforce needs and current performance.
  • Define measurable learning and performance objectives.
  • Design training activities around real workplace situations.
  • Develop resources, simulations, and opportunities for practice.
  • Implement the program in manageable stages.
  • Evaluate whether employees are applying the skills effectively.

However, the program should not focus only on technical proficiency. It should also examine the human qualities that determine how someone responds to change, pressure, and other people.

Questions should explore areas such as:

  • How has the employee responded to difficult situations?
  • What mistakes have they made, and what changed afterward?
  • How do they react when receiving feedback?
  • How do they treat coworkers whose roles carry less authority?
  • Do they listen before responding?
  • Do they take responsibility for shared spaces and outcomes?
  • Do they contribute to trust or participate in gossip and unnecessary conflict?

These questions provide a fuller picture of the individual and how that person is likely to interact with others.

The purpose is not to divide employees into those who are valuable and those who are not. Instead, the assessment can identify who is ready to model certain behaviors, who may need additional support, and which learning experiences will be most useful for each person.

A strong retraining program should develop both technical competence and human judgment. Technology, job requirements, and organizational structures will continue to change. The workers best prepared for that future will be those who understand themselves, listen to others, persist through difficulty, solve problems, and turn data into responsible action.

Why Is Gamification Harder to Implement in K–12 Education?

If gamification has become common in corporate training, why does it still seem difficult to implement consistently in K–12 education?

Rebecca Torchia of EdTech Magazine makes an important distinction between gamification and game-based learning. Gamification involves adding game-like elements to an activity that is not inherently a game, while game-based learning places the educational content inside an actual game structure.

Before reading about the topic, I did not realize there was a meaningful difference between the two. I do not come from a strong gaming background, and I have not been directly involved with many of the digital game-based approaches now used in K–12 classrooms. Once I understood the distinction, however, I could see why full game-based learning might be difficult for schools to adopt.

One challenge is public perception. The idea of students attending school to “play games” may not sit well with parents, taxpayers, or school board members, even when the activities are tied to learning objectives. I can imagine the headlines now: “Words with Friends Is the New Spelling Bee” or “Kahoot! More Like Give It the Boot!”

Another issue is personalization. Effective game-based learning may require more than a one-size-fits-all approach. Students differ in ability, motivation, prior knowledge, and comfort with competition. A game that challenges one student may frustrate another or fail to engage someone else. To be effective, the experience would likely need to be carefully designed and adjusted for different learners.

There is also the problem of time, workload, and distraction. Introducing a new platform or game structure means that both students and teachers must learn how to use it. If the process is overly complicated, the game itself can become another source of frustration. Teachers already manage lesson planning, assessment, classroom behavior, technology problems, and administrative responsibilities. Frankly, do they get paid enough to redesign every lesson as a game?

Gamification may be the easier starting point

Gamification, as Torchia describes it, may be easier to incorporate because competition and play are already natural parts of many learning environments.

My own K–12 experience provides a good example. The school district I attended was not particularly well funded, nor was it considered a highly rated educational system. Still, it did something well: teachers regularly added elements of gamesmanship to academic lessons.

I distinctly remember playing Around the World with math facts and shooting tennis balls into a trash can after answering questions about Texas history. Teachers took subjects that might otherwise have felt monotonous and made them more engaging through competition.

Those Friday activities turned the end of the school week into something resembling a play day, but we were still learning.

Instead of completing another pencil-and-paper quiz, we competed against our classmates in a game-show-style format to see who would advance to the next round. The game did not replace the lesson. It changed how we practiced and demonstrated what we had learned.

That same idea can still be applied with or without technology. Game elements can include points, levels, challenges, teams, progress indicators, rewards, or friendly competition. The instructor can choose what fits the learning objective and the students involved.

From a practical standpoint, this may also be easier to explain to parents and school boards than a proposal for full game-based learning. Adding game-like elements to a lesson sounds like an instructional strategy. Saying that students will learn through games may create more resistance, even when the underlying purpose is the same.

The role of productive discomfort

Jeanne Baker, founder of Brain Bakery, has argued that when a learner’s comfort level decreases, the eagerness to learn can increase. The idea is not to make students anxious or overwhelmed. It is to create enough challenge that they become curious, attentive, and motivated to improve.

Gamification can support that balance. A student may feel some pressure when answering a question in front of peers or competing to advance, but the playful structure can reduce the apprehension associated with a traditional test.

Shooting hoops with a tennis ball and participating in game-show-style quizzes were simple examples of this principle. The activities introduced challenge while lowering some of the tension surrounding academic performance.

Looking back, maybe our teachers knew exactly what they were doing.

Why corporate training may have an advantage

Corporate training often has fewer barriers because the learning goals are narrower and more directly connected to job performance. Employees may complete a gamified module to practice a specific process, learn a product, improve compliance, or develop a defined skill.

Organizations may also have more flexibility in selecting platforms, offering incentives, and measuring performance against a business outcome. In K–12 education, teachers must work within curriculum standards, testing requirements, accessibility obligations, limited budgets, and the expectations of families and governing boards.

Corporate learners are also adults who can usually understand why a game-like activity is being used. Younger students may focus more on winning than learning unless the activity is carefully designed.

Final thought

The difficulty is not that gamification has no place in K–12 education. It is that schools must balance engagement with learning objectives, equity, teacher workload, public perception, and instructional value.

Game-based learning may require substantial resources and design expertise. Gamification, however, can begin with something much simpler: a challenge, a point system, a team activity, or a tennis ball and a trash can.

The technology is optional. The intention is what matters.

The central question is not whether students should be allowed to play in school. It is whether play can be designed in a way that helps them participate, persist, and learn more effectively.