I used to think “AI in teaching” would be like buying a robot vacuum: press a button, sip coffee, and watch the magic happen.
Instead, it’s more like adopting a very fast puppy that can type. It’s enthusiastic, occasionally brilliant, sometimes confidently wrong,
and absolutely will chew on your deadlines if you don’t set boundaries.

The good news: when you build a working relationship with AIone grounded in your expertise, your course goals, and your students’ real needs
it can genuinely lighten the load. The better news: you don’t have to hand over the keys to do it.
This article breaks down how to partner with AI in a way that keeps learning human, keeps integrity intact, and keeps you in chargewhile also showing
where Cengage fits into the mix.

The Big Mindset Shift: AI Isn’t the ExpertIt’s the Assistant

Here’s the first rule of the relationship: your expertise is the steering wheel. AI can help draft, brainstorm, reorganize, simplify, and generate options,
but it can’t replace your judgment, your context, or your ability to say, “This doesn’t fit my students.”
When AI outputs are strong, it’s usually because a knowledgeable human gave it a clear goal, solid constraints, and then edited with a sharp eye.

Think of AI as a tool that amplifies what you already do well. If you’re clear about learning outcomes, you’ll get something usable faster.
If you’re fuzzy, you’ll get a beautifully formatted… pile of educational confetti.

Factor #1: Subject-Matter Expertise Still Does the Heavy Lifting

AI can produce fluent text, but fluency isn’t accuracy. Your subject knowledge is what prevents “sounds right” from becoming “is right.”
This matters most in three places: instructions, examples, and assessment.

Use AI for draftsthen apply your expertise like a quality filter

• Drafting rubrics: Ask AI for a rubric structure aligned to your outcomes, then revise descriptors so they match your standards and your course voice.
• Generating question banks: Have AI propose questions at multiple difficulty levels, then you verify correctness, remove ambiguity, and tune to what you actually taught.
• Creating examples and non-examples: AI can produce samples quickly; you refine them so they reflect common misconceptions you see in your class.

A practical reality check: if you wouldn’t copy-paste something from a random internet forum into your syllabus, don’t copy-paste AI output either.
The relationship works when AI accelerates your thinkingnot when it replaces it.

Factor #2: Clear Instructions Help EveryoneIncluding AI

If you’ve ever read a student submission and thought, “Oh no, they interpreted the assignment in a way I did not predict,” you already understand prompts.
Clear instructions aren’t just an AI thingthey’re a learning thing.

Many instructors are also designing for students with different processing styles and support needs.
When you write instructions that are specific, structured, and accessible, you reduce confusion, improve equity, and lower the chance that students use AI
as a panic button at 11:58 p.m.

A prompt formula that doesn’t feel like a tech bro handshake

1. Role: “Act as a writing coach / lab TA / statistics tutor.”
2. Goal: “Help me create three scaffolding activities for thesis development.”
3. Constraints: “No final answers. Ask guiding questions. Use plain language. Limit to 20 minutes.”
4. Context: “My students are first-year, mixed confidence, and we just covered X and Y.”
5. Output format: “Give me a table with time, activity, teacher script, and common pitfalls.”

Notice what’s missing: mystical incantations. You don’t need to “hack” AIyou need to communicate like the organized version of yourself
who exists briefly after coffee and before email.

Factor #3: Precision Instruction and Critical Thinking Can’t Be Outsourced

AI can generate possibilities. It can’t decide what matters. It can’t read the room. It can’t notice the class collectively blinking in confusion
at your example and silently screaming for a different one.

This is where teaching stays deeply human: you choose what to emphasize, what to revisit, what to slow down for, and when to challenge students
to wrestle with ambiguity. AI can support those decisions (suggesting alternate explanations, practice items, or discussion prompts),
but it doesn’t own them.

A simple standard: “Does this deepen thinking?”

Before using AI output in your course, ask:

• Does it reinforce the actual learning objectiveor just create busywork with better formatting?
• Does it encourage reasoning, reflection, and revisionor does it accidentally reward shortcuts?
• Would a student learn more by doing this, or by watching the AI do it for them?

Academic Integrity: Make It About Process, Not Just Policing

Generative AI didn’t invent academic integrity problems. It just made them faster and harder to detect with old-school methods.
If your policy is “Don’t,” but your assignment is “Write something generic that the internet has seen a million times,”
you’ve created the educational version of leaving a sandwich unattended in a room full of hungry teenagers.

A more durable approach is to design for transparency and learning processthen clearly communicate expectations.
Some instructors include explicit AI-use guidelines in syllabi and assignment descriptions, discuss ethical use early,
and emphasize that integrity is about demonstrating your learning journey.

The “traffic light” policy that students actually understand

• Green (allowed): brainstorming topics, outlining, practice quizzes, rewriting for clarity, generating study plans, asking for feedback on drafts (with disclosure if required).
• Yellow (ask first): summarizing sources, generating code, translating substantial passages, drafting lab reports, creating citationsanything that could blur authorship or accuracy.
• Red (not allowed): generating final submissions, fabricating data, writing discussion posts “as you,” or producing anything presented as original work without permission.

Pair the policy with assignments that show thinking: drafts, reflections, short oral explanations, in-class checkpoints,
or “explain your choices” memos. When students must document how they arrived at their work, AI becomes a toolnot a ghostwriter.

Where Cengage Fits: AI Built Around Course Materials (Not the Wild Web)

One reason AI in education can feel risky is that many general-purpose tools pull from broad internet patterns.
Cengage’s AI direction emphasizes using AI to support learning in ways that align with course content, teaching goals, and privacy needs.
In other words: less “random internet soup,” more “course-connected support.”

Cengage’s AI approach in plain English

• Personalized learning: helping students practice and close gaps without waiting for office hours.
• Enhanced teaching: supporting instruction with tools that complement classroom goals.
• Privacy and data security: applying safeguards and governance around education data.
• Human-centered design: building with faculty and student input, not in a vacuum.

The Student Assistant: guidance without just handing over answers

Cengage’s Student Assistant is positioned as a course-embedded helper that focuses on learning and critical thinking rather than simply outputting solutions.
The idea is to keep students working with trusted course resources and to nudge them through concepts step-by-stepmore coach than vending machine.

There’s also an Instructor Assistant concept aimed at giving instructors higher-level insight into where students are struggling,
so class time can target real friction points rather than guessing.

Practical Ways to Use AI Without Losing Your Soul (or Your Weekend)

Here are realistic, relationship-friendly ways to partner with AIespecially when combined with structured platforms and course materials.

1) Lesson planning that starts with your outcomes

• Ask AI to propose three activity options for one learning objective (discussion, practice, application).
• Choose one and adapt it to your discipline, your student context, and your course materials.
• Build in a short “show your thinking” checkpoint so students can’t autopilot.

2) Feedback that’s fasterbut still yours

• Paste your rubric descriptors (not student work) and ask AI to draft feedback sentence stems aligned to each criterion.
• Use those stems as a menu while you gradeediting so the feedback stays accurate, specific, and human.
• Reserve your time for comments that actually move the student forward.

3) Study support that encourages effort

• Have AI generate practice questions, then provide explanations that emphasize reasoning steps.
• Ask for “common wrong answers and why they happen” to teach misconception-busting.
• Use structured course tools for practice and retrieval, so studying is tied to what you assigned.

Privacy and Data: The Relationship Needs Boundaries

If AI is the puppy, student data is the couch you’d like to keep un-chewed.
The safest habit is simple: don’t paste sensitive or student-identifying information into tools that aren’t approved for that purpose.
Even well-intended uses (like running student work through an AI detector) can raise privacy and compliance concerns.

A good boundary is “data minimization”: use only what’s needed, anonymize when possible, and follow your institution’s guidance.
If your campus has approved tools or contracts that address data protection, use those instead of personal accounts.

The Human Parts AI Won’t Replace (And Why That’s a Relief)

AI doesn’t know your students’ lives. It doesn’t notice when a quiet student suddenly participates.
It can’t tell when a “wrong” answer is actually a smart insight expressed awkwardly.
And it can’t replace the trust students build when they feel seen.

That’s not anti-AI. That’s pro-teaching.
The healthiest working relationship is one where AI handles the repetitive tasks and your energy goes to the moments that change students.

A Simple “Working Relationship” Playbook You Can Actually Use

1. Pick one job for AI: brainstorming, drafting, question generation, or structureone at a time.
2. Set non-negotiables: your outcomes, your standards, your tone, your integrity rules.
3. Demand process: ask for reasoning steps, options, and promptsnot final answers.
4. Verify against course materials: if it’s not aligned, it doesn’t ship.
5. Edit like a professional: clarity, inclusivity, accuracy, and relevance.
6. Teach students how to use it ethically: model, practice, and reflectdon’t just warn.
7. Iterate: keep what helps learning; drop what creates shortcuts.

Common Pitfalls (So You Can Avoid Learning Them the Hard Way)

• Pitfall: Using AI to generate “final” course materials without review.
  Fix: Treat AI as a draft tool; you do the final pass.
• Pitfall: Assignments that reward generic output.
  Fix: Require personal application, reflection, drafts, or decision memos.
• Pitfall: Overreliance on detection tools.
  Fix: Design assessments for authenticity and be transparent about how evidence is evaluated.
• Pitfall: Privacy blind spots.
  Fix: Minimize data, follow institutional guidance, and use approved tools when available.

Conclusion: AI Can Be a Great CoworkerIf You’re the Manager

A working relationship with AI is less about finding the “best tool” and more about building the right habits:
clear goals, strong boundaries, thoughtful pedagogy, and human-centered judgment.
Cengage’s ecosystem adds a practical advantage when AI support is anchored to course materials and learning intentkeeping the relationship pointed
toward understanding, not shortcuts.

So yesAI can help. But don’t forget the most important part of this partnership:
you. You’re the one who knows what mastery looks like, what your students need, and what meaningful learning feels like.
AI is the assistant. You’re the educator. And that’s a power dynamic worth keeping.

My 7-Day “Working Relationship” Experiment with AI, Cengage, and Myself (Experience Add-On)

To make this concrete, here’s a realistic, composite “week in the life” of someone building a healthier partnership with AI and course tools.
Think of it as a field guide written by a human who occasionally forgets where they put their coffee.

Day 1: I stop asking AI to be a mind reader

Monday starts with the classic mistake: “Make this week’s lesson plan engaging.” AI responds with 47 ideas, 39 of which involve “small group discussion”
and one that appears to be a scavenger hunt designed by a raccoon. So I reset.
I give the real objective: “Students will compare two theories and justify which one applies to a case study.”
I ask for three options: one low prep, one medium prep, one high prep. Suddenly, the output is usable.
The relationship improves the moment I stop being vague.

Day 2: I use AI for structure, not authority

Tuesday is “rubric day,” otherwise known as “the day I consider becoming a lighthouse keeper.”
I ask AI to draft a rubric framework aligned to my criteria (claim, evidence, reasoning, clarity).
Then I rewrite the performance descriptors in my voice and align them to what students actually practiced.
AI saved me time on formatting and organization, but the quality came from the edit pass.
This is the sweet spot: AI accelerates the boring part; I protect the meaningful part.

Day 3: I design for process so integrity isn’t a guessing game

On Wednesday, I tweak an assignment to be more “process-forward.”
Instead of “Write an essay,” it becomes: submit a one-paragraph claim, a source grid, a draft, and a short reflection on what changed between drafts.
Students can still use AI in allowed ways (brainstorming, clarity, outlining), but they can’t skip thinking.
I’m not trying to trap anyoneI’m trying to make learning visible.

Day 4: I let course tools handle practice, and I handle teaching

Thursday is practice-heavy. This is where structured digital learning tools shine: students do targeted practice tied to the course materials,
and I review where they’re struggling. When AI support is embedded around the actual content students are expected to learn,
it feels less like “outsourcing the homework” and more like “getting coached through it.”
My role becomes diagnosing misconceptions and planning what to reteach, not re-explaining the same directions fifty times.

Day 5: I set privacy boundaries like an adult (finally)

Friday is the day I realize a lot of “helpful” AI habits can quietly become privacy problems.
I create a personal rule: no pasting identifiable student work into random tools.
If I’m using AI to generate feedback language, I feed it my rubric and a generic example I wrotenot a student submission.
If I’m tempted to use an AI detector, I pause and check institutional guidance first.
The relationship gets healthier when the boundaries are boring and consistent.

Day 6: I teach students how to use AI like a tool, not a personality

Saturday (yes, sometimes teaching brain leaks into weekends) becomes my “student-facing clarity” day.
I write a short AI-use guide: what’s allowed, what’s not, and what “good use” looks like.
I include examples like: “Ask for three thesis options and explain which one you chose and why,” versus “Write the whole assignment.”
Students don’t just need rules; they need models.
When they understand the difference between support and substitution, integrity stops being a mysterious vibe and becomes a shared practice.

Day 7: I reflect on what actually helped learning

Sunday is reflectionless “productivity guru” and more “what didn’t make me miserable.”
I keep the AI uses that supported clarity (drafting instructions, generating practice questions, outlining feedback stems).
I delete the ones that created noise (overlong brainstorms, generic activities, anything that didn’t match my outcomes).
The final lesson of the week is the simplest: the best AI workflow is the one that strengthens learning and gives you time back
without shrinking the human core of teaching.

That’s the working relationship: a practical partnership where AI supports the craft, Cengage tools anchor learning to course content,
and “myself” stays firmly in charge of the purpose, ethics, and judgment. The puppy can typebut you’re still the one running the house.

The post A Working Relationship with AI Technology, Cengage and Myself – The Cengage Blog appeared first on Global Travel Notes.

I used to be a proctored-exam believer. I liked the ritual: the quiet room, the “no phones” warnings, the collective sound of
keyboards clicking like a swarm of very polite crickets. Proctored exams felt official. They felt like education with a
badge and a flashlight.

Then online proctoring arrived, and the badge-and-flashlight vibe turned into “please rotate your webcam and prove you’re not
hiding answers behind your laundry basket.” Suddenly, exams weren’t just about what you knew. They were about whether your
internet behaved, whether your face looked “normal” to a camera, and whether you could sit still like a museum statue for two
hours straight.

Somewhere in that shift, I realized a hard truth: many proctored exams don’t actually measure learning. They measure
compliance under pressure. And compliance isn’t the same thing as competence. That’s why I moved from proctored
exams to open-book examson purpose, with a plan, and with fewer awkward webcam angles.

Proctored exams: what they’re good at (and what they’re not)

Let’s give proctored exams their credit. In certain contexts, they make sense. If a skill requires quick recall for safety or
real-time performancethink medication calculations, emergency procedures, or specific licensing requirementsthen a controlled
testing environment can be justified.

But in many everyday courses, proctoring becomes a stand-in for rigor. We start believing that if a test is locked down hard
enough, learning must be happening. That’s like assuming a restaurant is five-star because the menu is laminated.

The “control theater” problem

Proctoring often creates what I now call control theater: the illusion that we’ve secured honesty because we’ve
increased surveillance. The exam feels stricter, so it feels fairer. But fairness isn’t just about catching bad behavior. It’s
also about not punishing the people who are already at a disadvantagestudents with anxiety, students who test differently,
students whose technology is unreliable, and students who don’t have a quiet, private space to perform “perfect test-taking.”

Proctoring can amplify test anxiety

Test anxiety is real, common, and not distributed evenly. Add a proctorespecially a remote proctorand you can turn “normal
nerves” into “my brain has left the building.” Research on online proctoring has found that anxiety can interact with proctored
settings in ways that hurt performance for students who already score high on trait test anxiety.

That matters because the score you see at the end may reflect stress management more than mastery. If the goal is to measure
learning outcomes, we should be careful about exam conditions that systematically distort those outcomes.

Privacy, equity, and the tech-glitch tax

Remote proctoring doesn’t just “watch.” It can require permissions, recordings, room scans, identification checks, and behavior
flags. That raises predictable questions: Where does the data go? Who sees it? How long is it stored? What happens when the
system misreads a student’s movement, lighting, or assistive behavior as suspicious?

Even when everything works perfectly (a rare holiday miracle), students still pay a “tech-glitch tax”: the cognitive load of
worrying about battery life, wifi drops, background noise, roommates, family, pets, and whether the software will suddenly
decide their face looks too… face-like.

Open-book exams: what they actually test

“Open-book” sounds like a free pass until you design it correctly. A good open-book exam is not “What’s the definition of
photosynthesis?” while students sit next to a biology textbook the size of a small refrigerator.

A good open-book exam asks: Can you use what you know? Can you interpret information, connect concepts, justify
decisions, and solve problems in context? That’s the kind of thinking most workplaces rewardand it’s the kind of thinking that
memorization-heavy exams sometimes miss.

From “remembering” to “reasoning”

Traditional closed-book exams often prioritize recall because recall is easy to test quickly. But open-book formats push you to
assess higher-order skills: application, analysis, evaluation, and creation. When students can reference notes, the question has
to do more than ask for a fact. It has to ask for judgment.

For example, instead of asking, “What is supply and demand?” an open-book exam can ask:

• Case scenario: “A city caps rideshare prices during a major storm. Explain what happens to availability and
  wait times. Use supply-demand logic and address at least one unintended consequence.”
• Data interpretation: “Here’s a simple chart of price changes over time. Identify the likely cause and defend
  your reasoning with two concepts from the course.”
• Trade-off decision: “Recommend a policy option and explain which stakeholder it helps and which it harms,
  using evidence from your readings.”

Students can look up a definition, sure. But they can’t instantly look up your reasoning, your explanation, or
your ability to connect ideas under time constraints.

Open-book exams are closer to real life

Outside of school, most professionals don’t solve problems by locking their books in a drawer and racing a timer. They consult
references, collaborate, check assumptions, and document decisions. Open-book exams mirror that realityespecially in fields
like business, healthcare, tech, education, and policy.

In other words, open-book exams can be less about “Do you know it?” and more about “Do you know what to do with it?”

The big fear: “But won’t students just cheat?”

The fear is understandable. If you remove proctoring, you remove a visible deterrent. But here’s the switch that changed my
whole approach: integrity is mostly a design problem, not a surveillance problem.

If your questions can be answered by copying a sentence from a PDF, the issue isn’t that the exam is open-book. The issue is
that the exam is asking for something that doesn’t require thinking.

Design beats surveillance

When I began writing open-book exams, I stopped asking for “the right phrase” and started asking for “the right thinking.”
That meant:

• Using scenarios, mini-cases, and datasets that require interpretation.
• Asking students to explain why, not just choose what.
• Requiring a short justification, even for multiple-choice selections.
• Including “show your work” or “explain your approach” prompts where appropriate.
• Asking students to compare two options and defend a choice using course concepts.

The result: the exam became harder to outsource and easier to grade for genuine understanding.

Practical safeguards that don’t feel like a spy movie

Open-book doesn’t mean open-chaos. I still set boundaries that support fairness:

• Clear rules: What resources are allowed (notes, textbook, course slides, etc.).
• Time limits: Enough time to think, not enough time to “research a whole new personality.”
• Question variety: Different versions, shuffled order, and question banks when available.
• Unique prompts: Context-specific questions that don’t map neatly onto a search result.
• Reflection items: “What concept was most useful here and why?” (Hard to fake convincingly at scale.)

This isn’t about outsmarting students. It’s about aligning the exam with what you truly want to measure.

What students gain (besides the right to blink)

Less anxiety, better thinking

When students believe the exam is about reasoningnot perfect recallthey often prepare differently. Instead of memorizing,
they organize, practice applying concepts, and learn how to find and use information efficiently.

Many students report that open-book formats feel more fair because they reward preparation and understanding. And when anxiety
drops, performance can become a clearer signal of learning.

More accessible assessment

Open-book exams can reduce barriers for students who struggle in high-surveillance settings. They also lessen the impact of
“environment inequality”not everyone has a quiet, private room with perfect lighting and zero background noise.

Accessibility still requires thoughtful design (and support services where appropriate), but moving away from invasive proctoring
can remove a major friction point for many learners.

Better alignment with modern learning (and modern tools)

In a world where information is available instantly, education can’t only be about storing facts. It has to be about using facts:
evaluating sources, applying concepts, and communicating reasoning. Open-book exams push instruction in that direction.

When I still consider proctored exams

I’m not anti-proctoring in every scenario. I’m anti-proctoring-as-default. I still consider controlled assessments when:

• Safety requires recall (for example, high-stakes clinical or lab procedures).
• Accreditation or licensing standards require specific exam conditions.
• Baseline skills must be demonstrated without aids (like foundational fluency before advanced work).

Even then, I ask: can we proctor less and assess more? Sometimes the best answer is a mix: smaller proctored
checks paired with open-resource assessments that measure applied competence.

My mini playbook for switching to open-book

1) Decide what “open-book” means in your context

There’s a spectrum:

• Open-note: students use their own notes.
• Open-book: notes + textbook.
• Open-resource: course materials, handouts, approved references.
• Open-web: broader internet access (risky unless questions demand deep reasoning).

The more open the resources, the more your questions should emphasize analysis and explanation.

2) Write questions that require thinking, not hunting

A quick test I use: if a student can answer by copying the first paragraph of a search result, the question is too easy for an
open-book format. Add a scenario, add constraints, ask for justification, or require a comparison.

3) Make expectations painfully clear (in a kind way)

Students do better when they know what success looks like. I tell them:

• What resources are allowed (and what aren’t).
• Whether collaboration is permitted.
• How answers will be graded (especially explanations and reasoning).
• What academic integrity means for this assessment.

4) Practice the format before the high-stakes moment

If the first open-book exam is the final, students may treat it like a scavenger hunt. Low-stakes practice helps them learn how
to prepare: organize notes, make quick-reference summaries, and practice applying concepts under time limits.

The bottom line

I switched to open-book exams because I wanted assessment to feel more like learning and less like surveillance. Proctored exams
can sometimes be necessary, but they’re too often used as a default settinglike the password “1234,” except with more stress.

Open-book exams, when designed well, reward understanding, reduce unnecessary anxiety, and reflect how people actually solve
problems in the real world. The rigor doesn’t disappearit moves. It shifts from “Can you memorize under pressure?” to “Can you
think, decide, and explain?”

And honestly? That’s the kind of education I want to be part of.

Extra: of Real-Life Experience (a.k.a. how this change felt in practice)

My first serious break-up with proctored exams happened during a remote test that should’ve been simple. I was ready. I had
studied. I had water. I had snacks. I had the confidence of someone who had highlighted exactly three lines in the textbook and
thought, “Yes, this is knowledge now.”

Then the proctoring software started its pre-exam routine. It wanted permissions. It wanted camera access. It wanted my browser
to behave like a locked door. It wanted me to prove I wasn’t hiding answers in the general vicinity of my desk, my walls, or
apparently my soul. By the time the exam began, my brain had already spent half its energy on logistics and worry.

Midway through, I became intensely aware of my own eyeballs. Not in a philosophical waymore like, “Am I blinking too much?”
It’s hard to focus on a complex question when you’re also trying to look like a perfectly innocent human being who has never
once thought a suspicious thought. Add a tiny internet hiccup and the fear of being flagged, and you get the academic version
of trying to parallel park while someone films you for a documentary called Driver Under Pressure.

When I later tried an open-book format, the vibe changed immediately. I still had to prepare, but I prepared differently. I
stopped cramming definitions and started building a “concept map” in my notes: key ideas, how they connect, what assumptions
they depend on, and common mistakes. I practiced explaining concepts in my own words, because I knew the exam would ask for
reasoning, not trivia.

The exam itself felt more like problem-solving and less like a hostage negotiation with my webcam. I could look up a detail if
I truly needed itbut I didn’t have time to “research.” That’s the funny secret of a well-timed open-book exam: it rewards the
student who already understands where the information fits. If you don’t understand, having a book nearby is like having a
fire hose when you need a glass of watertechnically helpful, but mostly chaotic.

I also noticed something I didn’t expect: open-book exams exposed shallow learning more clearly. If a student had only
memorized phrases, the moment the question demanded application, the memorization fell apart. Meanwhile, students who had built
real understanding could explain, compare, and defend their answers with confidenceeven if their wording wasn’t perfect.

Over time, the switch improved my teaching and my trust. Instead of designing assessments to “catch” students, I designed them
to challenge students. The tone shifted from suspicion to scholarship. And in a world already full of surveillance,
that shift felt not just educationalbut humane.

The post Why I Went from Proctored Exams to Open-book appeared first on Global Travel Notes.