Learning by Doing, at Scale: Cambridge’s Integrated Approach to Artificial Intelligence

How an AI-forward healthcare college uses technology to strengthen — not replace — hands-on student development

Ask ten colleges how they are approaching artificial intelligence and you will often get ten versions of the same answer: a new elective, a policy on academic honesty, and a workshop or two. Useful, but narrow. Artificial intelligence is not showing up in healthcare as one job title or one software category. It is showing up inside the electronic health record, inside coding and claims workflows, inside imaging queues, inside scheduling and staffing systems, and inside the security tools that protect patient data.

An education that treats AI as one course teaches students about a tool. An education that integrates AI teaches students to work in an environment where the tool is already present. Cambridge College of Healthcare & Technology has taken the second path, and the reasoning behind it traces back to a well-documented finding about how people actually learn.

What is the “doer effect,” and why does it shape how Cambridge teaches?

The doer effect is a learning-science principle established by researchers at Carnegie Mellon University’s Open Learning Initiative. Studying data across four courses and more than 12,500 students, the researchers found that completing interactive practice activities had about six times the effect on learning outcomes that reading the same material did. Follow-up work established that the relationship is causal: doing the practice produces the learning. It is not simply that stronger students happen to do more exercises.

Two details matter for a career-focused healthcare college. First, the effect holds across levels of prior knowledge, meaning it benefits students who arrive underprepared as much as — often more than — those who arrive confident. Second, students consistently underestimate it. Given a choice, most learners will reread a chapter rather than test themselves on it, because rereading feels productive while practice feels difficult. The feeling is misleading. The difficulty is the learning.

This finding did not change Cambridge’s philosophy so much as confirm it. A nursing student does not become competent by reading about sterile technique. A medical billing student does not learn claim denials by memorizing a code set. A cybersecurity student does not learn incident response from a lecture. Cambridge was built around skills labs, simulation, and supervised clinical practice long before anyone described the underlying principle in a journal.

The open question was never whether practice works. It was how to give every student more of it.

How does AI let Cambridge scale learning by doing?

Practice has always been the expensive part of education. A well-written practice question takes time to build. Meaningful feedback takes an instructor’s attention. A realistic scenario takes design work. Those constraints are the reason so many courses drift toward reading and lecture: they scale, even though they teach less.

This is where AI earns its place in the curriculum. Used carefully, it lowers the cost of the thing that works. It can generate practice questions tied to a specific reading, so students test themselves as they go instead of at the end. It can return feedback in the moment rather than a week later. It can produce variations on a case scenario so a student sees the same clinical reasoning problem from six angles instead of one. It can surface a pattern in a student’s errors that neither the student nor the instructor would have spotted from a single grade.

None of that replaces the instructor. It replaces the silence between a student’s question and the next class meeting — which, for working adults studying at night, is often where learning stalls out entirely.

Where does AI appear in Cambridge’s nursing programs?

Nursing is where the integration has to be most careful, because the stakes are highest and the hands-on requirement is non-negotiable. Across Cambridge’s nursing pathways — from the Nursing Assistant and Practical Nursing diplomas through the Associate of Science in Nursing, the RN-to-BSN completion track, and the MSN Family Nurse Practitioner program — AI supports preparation and reasoning practice, while competency is still demonstrated in the lab and at the bedside.

In practice, that means students use AI-supported practice to rehearse before simulation rather than during it: working through patient scenarios, drilling pharmacology, explaining their clinical reasoning and having it questioned, and preparing for NCLEX-style testing with unlimited repetitions rather than a fixed question bank. A student who has already reasoned through twenty variations of a deteriorating patient arrives at simulation ready to perform, not ready to guess.

Students also learn about the AI already embedded in the environments where they will work: sepsis prediction and early-warning scores, fall-risk and readmission models, documentation assistance, and acuity-based staffing tools. A nurse who understands why an early-warning algorithm fired — and, just as importantly, when to escalate despite it staying silent — is a safer nurse. That judgment is a clinical skill now, and it is taught as one.

Where does AI appear in health information and revenue cycle programs?

Healthcare data is where AI is changing daily work fastest. Cambridge’s medical billing and coding, health information technology, and health information management programs now assume that automated coding suggestions, computer-assisted documentation, and predictive denial management are part of the environment, not a future development.

That shifts what the job actually requires. The coder who competes on speed alone is competing against software. The coder who can audit an AI-generated code assignment, identify where the documentation does not support it, and defend the correction to a payer is doing work that automation cannot replace — and is doing it at a higher level. Cambridge students practice exactly that: reviewing machine output, finding the error, and articulating why.

The same logic runs up the ladder into health informatics, where graduates work on the governance side — data quality, model monitoring, and the question of whether a tool performs equitably across the patient population it serves.

Where does AI appear in cybersecurity and IT programs?

In healthcare cybersecurity and information technology, AI cuts both directions, and students study both. Defensively, machine learning drives anomaly detection, alert triage, and behavioral analytics. Offensively, generative tools have made phishing more convincing, social engineering more scalable, and reconnaissance faster.

Students practice on both sides of that line: tuning detection logic, triaging alerts and reducing false positives, and analyzing AI-assisted attack patterns in a controlled environment. Because Cambridge’s cybersecurity programs are healthcare-anchored, the coursework also covers the constraints unique to clinical settings — medical devices that cannot be patched on a normal schedule, protected health information under HIPAA, and the operational reality that a hospital cannot simply take a system offline during patient care.

How do Cambridge faculty use AI themselves?

An integrated approach fails if the faculty are teaching from the outside. Cambridge instructors use AI in their own preparation — building practice sets, generating scenario variations, reviewing where a cohort is struggling, and drafting materials they then revise with their own clinical judgment. Because most Cambridge faculty come from active practice, they bring back what they see in the field, including the failures.

That last part matters more than it sounds. Faculty who use these tools regularly know where they break: the plausible answer that is wrong, the citation that does not exist, the recommendation that ignores a patient’s comorbidity. An instructor who has been burned by an AI tool teaches skepticism far better than a policy document does.

Cambridge also treats faculty development in AI as ongoing rather than one-time. The tools change on a timeline measured in months, and a curriculum that is refreshed every five years cannot keep pace with that.

Does using AI mean students learn less?

It can — and this is the risk Cambridge takes seriously. The same research that makes the case for AI-supported practice also explains how AI can undermine learning: if a tool hands a student the answer, the student stops doing the practice, and the six-times advantage disappears. Worse, the student feels like they learned, because fluent output produces a convincing illusion of understanding.

Cambridge’s answer is structural rather than punitive. Assessment is anchored where AI cannot stand in for the student: skills checkoffs, simulation performance, supervised clinical hours, hands-on labs, and externship evaluations. A student can generate an essay about wound care. A student cannot generate competent wound care in front of a clinical instructor.

Alongside that, students are taught the discipline that actually transfers to the workplace: verify before you trust, know what the tool cannot see, and never sign your name to output you cannot defend. In healthcare, accountability does not transfer to software. The credential holder remains responsible.

What does an AI-integrated education mean for your career?

Healthcare employers are moving faster than most curricula. Hiring managers in nursing, revenue cycle, informatics, and security increasingly ask candidates how they have worked alongside these tools — and the answer that lands is specific, not general. “I have used an AI-assisted coding tool and I know how to audit its output” is a different answer than “I am comfortable with technology.”

Integration is what makes the specific answer possible. A student who has used these tools inside pharmacology, inside coding practice, inside alert triage, and inside their clinical preparation has accumulated real examples. A student who took one AI elective has a line on a transcript.

Cambridge’s stackable credential ladders reinforce that over time. Because programs are designed to build — diploma to associate to bachelor’s to master’s — a graduate can return as the field shifts without starting over. Details on Cambridge’s online program options are available for students balancing school with work and family.

What stays the same at Cambridge?

Everything that matters most. Students still put their hands on equipment. They still practice on manikins before patients and on patients under supervision. They still complete clinical rotations and externships at healthcare sites across Florida and Georgia — relationships built over years that no national online program can replicate from a distance. They still learn from instructors who have done the work.

The doer effect research points in exactly this direction. If doing is what produces learning, then a college organized around doing was already aligned with the evidence. AI is a way to give students more repetitions, faster feedback, and better preparation for the moments that count. It is a multiplier on the model. It is not a substitute for it.

Frequently Asked Questions

Does Cambridge allow students to use AI in their coursework?

Cambridge teaches students to use AI the way professionals do — as a tool that supports thinking rather than replaces it. Specific expectations vary by course and are set by the instructor, and competency is always demonstrated through hands-on assessment.

Do I need a technical background to succeed in an AI-integrated program?

No. Programs are designed for students entering from a range of backgrounds, including career changers with no prior healthcare or IT experience. AI concepts are introduced in the context of the work rather than as standalone technical theory.

Will AI replace the jobs I am training for?

The clearer pattern is that AI is changing roles rather than eliminating them, and shifting value toward judgment, verification, and patient interaction. Roles that involve hands-on care, oversight of automated output, and accountability for decisions are the ones this training targets.

Is this available online?

Many Cambridge programs are offered online or in a hybrid format, with hands-on and clinical components completed at approved sites. See Cambridge online programs for current offerings and format details.

Does Cambridge offer a degree specifically in healthcare AI?

Cambridge continues to expand its credential ladders in response to healthcare workforce demand. Prospective students should contact an admissions advisor for current program availability.

How is this different from a general AI course?

A general course teaches about AI. An integrated approach teaches students to work with AI inside the actual tasks of their profession — clinical reasoning, coding and claims, alert triage — which is how the tools appear on the job.

Source note: The doer effect research referenced in this article originates with Kenneth R. Koedinger and colleagues at Carnegie Mellon University’s Open Learning Initiative, including findings published in Learning Analytics & Knowledge.