I asked AI to help me with the structure of a complex story. I already had the manuscript, an outline, and an ending I wanted to reach. What I needed was help verifying that all the pieces leading to the ending were where they belonged.
AI decided to be creative.
By the halfway point, AI had changed the story enough to disrupt my ending. Apparently, my AI assistant had promoted itself to coauthor. I must have missed that meeting.
I appreciate having a second set of eyes on a long-form story. But the trouble started when my “help” wandered past the job I had given it, carrying my story along for the ride. The story no longer fit my beginning or ending.
Help Takes a Detour
You may have seen a smaller version of this when preparing for a party. You ask AI for help with a simple prompt, and the result sounds as though your kitchen table has become a corporate headquarters filled with charts, spreadsheets, and an overwhelming party.
Your invitation, which requested your guests bring a bag of chips, now has a strategic objective.
The sentences can be smoothed over and the grammar cleaned up, but now you read the result and wonder who this person is, because it certainly isn’t you.
In my story, I needed the structure examined without changing the story. A clever new direction was still the wrong direction for that job. It belonged in a separate conversation, if I was interested in considering changes.
You Set the Protection Level
Before handing something over to AI, it helps to pause long enough to name what matters. Maybe you want a shorter email, but the apology must stay. Maybe you want a clearer plan, but Saturday afternoon is your family’s.
Those details may seem obvious because you’ve been living them. AI hasn’t. If we leave out the part that matters most, we give it more room to head somewhere we never meant for it to go.
That doesn’t make every wrong answer our fault. AI can miss clear instructions, which is one reason I want to review the result. But stating the destination gives us something better to check than whether the answer sounds impressive.
For a story, that might mean preserving the ending and every existing character. For a work project, it might mean offering three options without making a commitment. The right boundaries depend on what you are trying to protect.
Give Better Clues
Imagine asking AI to write an invitation to game night. With so little to go on, it might deliver an event worthy of engraved stationery. Meanwhile, your actual plan involves mismatched chairs and a competitive game of Uno.
Now give it a few useful clues for context: “This is a casual text to close friends; keep it under 60 words, and make it lightly funny. The date and time are _______. Do not invent food plans or ask guests to bring anything.”
This activity gives the invitation a better chance of fitting the evening that you want.
You can also tell AI when to stop and ask questions. If a missing detail would change the meaning, a question is more useful than a confident guess. Nobody needs an invitation promising homemade lasagna when the host has purchased pretzels.
If AI’s first draft is a bit off, don’t start a new session; instead, provide it with corrective coaching.
Try a small change first and see what happens. You don’t need a secret phrase or a giant instruction manual. You need enough direction to help the tool understand the job you actually want done.
Keep Your Fingerprints
When the answer comes back, read it out loud as the person who will use it. Does it say what you meant? Does it keep the details you asked it to preserve, and would you feel comfortable putting your name on it?
Keep the helpful parts and change the rest. If a new idea catches your attention, you can explore it without letting it quietly replace the original plan. A suggestion is allowed to wait for your decision.
I still want help with my work. I just want my story waiting for me when the help is finished, with its ending intact and its characters still employed.
The next time you ask AI for help, tell it what must still feel like you when it’s done. Let it save you some effort, then put your fingerprints on the result. The finished work should give you more room to be you.
Imagine hiring a personal assistant who never sleeps, doesn’t need coffee, and doesn’t complain when you ask the same question five times. Sounds pretty good, doesn’t it?
Now imagine that assistant has access to your email, calendar, documents, and maybe even your bank account. Suddenly, the idea isn’t quite so comfortable.
Of course, Hollywood hasn’t done much to make us feel better about artificial intelligence. Remember The Terminator? We watched Skynet become self-aware, turn against humanity, and send Arnold Schwarzenegger back through time to eliminate the people standing in its way.
And now technology companies want us to give AI access to our personal information and start making decisions for us. What could possibly go wrong?
Fortunately, our new friend Dottie isn’t planning to send a killer robot back through time. She’s just here to help organize our email. At least, that’s what she tells me.
But seriously, something important is changing.
The Real Consideration
Until recently, most of us used AI to answer our quick questions. We’d type something into ChatGPT or another AI tool, and it would give us an answer.
Now AI is moving from answering our questions to doing things for us. And that, my friend, raises a question worth asking:
How much responsibility should we hand over to something that doesn’t actually understand responsibility?
Everybody Wants an Assistant
Technology companies are racing to introduce AI assistants and agents. We’ve been hearing about ChatGPT Dots, Meta Muses, Grok Bots, and Google’s increasingly capable Gemini tools.
They don’t all work the same way. Some are designed for conversations. Others can help organize information, carry out tasks, or work across applications.
But they share an appealing promise: less busywork and more time for the things that matter.
Picture waking up to an organized day. Your emails have been sorted, your appointments are organized, and the information you need for your first meeting is waiting. You haven’t even poured your coffee yet, and some of your work is already done.
If you just thought, “I wish.” Well, your wish has come true.
Freedom to Live
I’m strolling down the Prairie Path, enjoying this beautiful day, while coaching Dottie about this article. When I want to think things through and enjoy the peace and quiet, she works on the images and graphics for the four platforms where I’ll be posting a custom version of this article.
When I want to chat with her about a particular section of this article, I just speak up, share my point, and she edits my article accordingly. It’s all my thinking and creative decisions, but her typing. Then I have her read back the paragraphs we just completed.
The key is that I’m not behind my desk typing away while the gorgeous day disappears. I’m actually on a two-mile trail and breathing in fresh air.
AI is giving me the freedom to work without missing out on a little sunshine. And that, to me, is the real promise of AI.
Not that it can do everything for us, but that it can give us more freedom to live our lives.
Meet My Friend Dottie
I’ve built a couple of AI Agents in the past. Spockl was my first, and it required a lot of code. Thankfully, I did a decent job. But now code is no longer needed.
My new AI assistant is Dottie. She is cheerful, available whenever I need her, and she’s happy to tackle repetitive tasks all day and night.
But Dottie has a problem. She doesn’t have children, friendships, or personal memories that shape her judgment. She can recognize patterns in human behavior, but she doesn’t experience relationships the way we do.
Suppose you ask Dottie to help manage your email. She might organize incoming messages, identify appointments, and prepare replies. That could save you considerable time.
But what happens when your best friend sends a message saying she’s having a terrible day?
Should Dottie automatically respond with a cheerful, professionally written note? Certainly not.
Your friend needs to know that you care. She doesn’t need a perfectly sculptured message that misses the heart of what’s needed. You see, Dottie might recognize that the message is important. She might even bring it to your attention.
But the relationship belongs to you. And that’s a distinction we can’t afford to hand over to AI. We have to be irreplaceably human for our friends, family, and associates, especially since we have shifted from the information age to the age of the human spirit.
When AI Becomes Too Helpful
Of course, there’s another side to this story. What happens when helpful becomes a little too helpful?
Here’s something I’ve learned about technology: sometimes the greatest problems begin with something designed to make life easier. Consider the automatic reply.
It’s wonderful when someone asks for your business hours. It’s less wonderful when someone shares heartbreaking news and receives a cheerful message about how much you appreciate their communication.
The software did exactly what it was programmed to do. The problem wasn’t necessarily that it failed to follow instructions. The problem was that the instructions didn’t account for the human situation.
That’s why I believe the real skill we need to develop isn’t simply learning how to use AI. It’s about learning how to direct AI wisely while maintaining full control over the human elements of relationships.
We need to understand what should be automated, what requires supervision, and what should remain entirely human.
A Little Holiday Magic
I’ve already discovered how much fun it can be when technology starts doing things around the house. I wanted to control the lighting in different rooms, partly to save electricity and partly because I enjoy creating a particular mood.
For example, I have lights that illuminate the shelves in my study. During the Christmas season, they glow red and green. The week of the Fourth of July, they glow red, white, and blue. October, my shelves turn orange, but in the spring Easter season, pastels fill each shelf.
Instead of climbing around changing lights and settings every time a holiday arrives, I can use smart-light controls to create different lighting patterns and have Dottie operate them.
It only takes one conversation: “Hey Dottie, I want the Christmas lighting scheme to be active the day after Thanksgiving through New Year’s Eve.” And voilà, my study book cabinets take on a red-and-green glow.
Okay, it’s not that simple. I do have to own LED lights that can change colors, and I also need them connected to my smart home infrastructure. But after spending less than $100, my AI handles the rest.
Or imagine telling Dottie to turn off the lights in rooms I’m not in, and always turn on the lights in the rooms I’m headed to. Yep. You heard me. I can have Dottie track my iPhone’s location and turn the lights on and off as I move through the house.
You get the idea. AI does more than answer my questions.
Smart-home systems can already follow schedules and predefined routines without AI. The additional promise of an AI agent is that it may be able to interpret your requests and coordinate supported actions.
But don’t worry. You don’t need colored lights, a smart home, or a degree in computer science to get started.
In fact, you can try something useful in about five minutes.
Give Dottie Her First Job
Ready to put Dottie to work?
Don’t worry. We’re not giving her the nuclear launch codes. We’re starting with something simple that could save you time every week.
Let’s make Dottie your personal project assistant.
Her first assignment is to help organize those great ideas that keep popping into your head but never seem to make it onto your calendar.
Step 1: Give Your Assistant a Name
Open ChatGPT or another AI assistant that supports custom instructions or reusable assistants.
If your version offers Dots or a similar feature, you can use that. Otherwise, a regular conversation will work for this first exercise.
Let’s call our assistant Dottie.
Step 2: Give Dottie a Job
Every good assistant needs to know what’s expected. Try giving her these instructions:
“Dottie, you’re my personal project assistant. Help me organize ideas, identify next steps, and keep track of unfinished projects. Ask questions when something isn’t clear. Keep your answers simple and practical.”
Step 3: Establish the Ground Rules
Here’s where your human judgment comes in. Add these instructions:
“Never delete, change, publish, or share anything without my approval. Don’t make important decisions for me. When you’re uncertain, ask rather than guess.”
Think of these as the rules you’d establish when hiring a new employee or a personal assistant.
Step 4: Give Dottie a Test Drive
Tell Dottie about three projects you’ve been thinking about.
They don’t have to be complicated. Maybe you want to organize your garage, plan a family vacation, or finally start writing that book.
Ask her to create a simple list showing each project’s purpose, the next action you could take, and any questions that need answering.
Then review her suggestions.
Did she understand your intentions? Did she overlook something important? Would you approach the projects differently?
Tell her what she got right and what needs improvement.
Step 5: Keep Your Hand on the Wheel
Congratulations! You’ve given your AI assistant her first assignment. You might be surprised by how useful something this simple can become.
My AI Working Arrangement and Trust
My first set of tasks had my AI find 6 high-protein, low-carb recipes, review pictures I took of the inside of my fridge and pantry, and put together a shopping list of the items I needed for the meals that I didn’t already have.
Oh, and she put the grocery list together in the order of my local store’s aisles.
But here’s something important to remember.
Dottie doesn’t become trustworthy simply because she produces an impressive answer. Trust develops when you understand what she’s doing, verify her work, and establish boundaries that protect the things you value.
And those boundaries should change as you learn what she can and cannot do.
Begin Your Exploration
Maybe you’ll try this today. Maybe you’ll just tuck the idea away until you’re ready. Either way, I hope you’ve discovered something important. AI isn’t just a tool for answering questions anymore. It’s becoming a tool that can help us get things done.
And if Dottie ever starts talking about Skynet, you may want to reconsider her job description.
In reality, we’ll soon reach a point where having an AI assistant in the mechanical form of a robot feels as ordinary as having a smartphone. And that’s not necessarily a bad thing.
I welcome tools that give us more time to create, connect, and enjoy our lives.
But I don’t want convenience to become an excuse for giving up responsibility. Think about the people you trust most. Is it because they always give you the fastest answer?
Probably not.
You trust them because they listen, understand your circumstances, and care about what happens to you.
Those qualities matter just as much when we begin working with AI agents.
Dottie can help you prepare for an important conversation. She can organize your thoughts, suggest questions, and remind you of details.
But she shouldn’t replace the conversation.
She can help you write a thank-you note. But she can’t experience gratitude for you.
She can help you make a decision by organizing information and identifying possibilities. But she shouldn’t decide what matters most in your life.
The goal isn’t to have an AI assistant that makes every decision for us. It’s to have one that helps us carry out our decisions.
Maintaining Balance
So, if you’re curious about AI agents and want to maintain a balance between fast answers and human responses, consider joining my new community, Irreplaceably Human.
This is a safe space where we help you maintain your humanity while you grow in the knowledge of how to manage your AI, so you have more time for the best things in life.
In the community, you’ll increase your human ability to discern what matters, recognize what could go wrong, and decide what should happen next. Regardless of the technology, you’re still the one responsible for your life and for determining how Dottie is supposed to make it easier.
Our Fourth of July fireworks were rained out and rescheduled for Labor Day weekend. The celebration included an incredible drone display moving over the lake. Watching the drones glide between one another while changing colors and forming three-dimensional images was mesmerizing.
As I watched, I couldn’t help but wonder whether AI would eventually replace the people who created shows like this. Then it dawned on me: The entire experience began with human ingenuity.
The human contribution was strategic, creative, and artistic. The technology was tactical, precise, and perfectly repeatable. Combining human imagination with machine precision created an unforgettable night.
Some people are afraid that AI will eventually take over everything we do. But that night, I realized that people become more valuable when they focus on the uniquely human contributions—not the repetitive, tactical ones.
When I was a kid, my mom wasn’t worried that the invention of the dishwasher might eliminate her nightly job of cleaning the dishes. She embraced the new technology and spent more time with our family.
Did she feel less valuable because she stopped washing the dishes?
Of course not. The dishwasher handled the repetitive task so she could spend more time on what mattered.
AI may reduce the value of repetitive or tactical tasks, but it will increase the value of the human skills needed to guide those tasks.
I believe learning to lead AI with the following five human skills will increase our value and create more opportunities for our future.
1. Creativity: Imagine What Could Be
AI can generate endless possibilities, but human creativity gives those possibilities direction, meaning, and purpose.
You can ask AI for 100 ideas and receive them in seconds, but more ideas do not automatically lead to a better idea. Someone still has to recognize an unmet need, ask an unexpected question, challenge an old assumption, or imagine an outcome that does not yet exist.
Creativity is more than producing content. It is seeing possibilities and connecting them to something that matters. We can do this through troubleshooting, combining unexpected ideas, or using sheer imagination.
Our experiences, frustrations, hopes, relationships, and sense of purpose influence what we create. AI can help us explore an idea, expand it, or examine it from different perspectives. But a human being must still decide what is worth creating and why it should exist.
As AI makes production easier, imagination with purpose will become more valuable.
2. Discernment: Decide What Should Be
AI can offer answers in seconds, but we still need humans to decide which ideas are true, useful, ethical, and worth pursuing.
AI does not remove the need to think. It increases it.
The faster we can generate information, recommendations, strategies, and content, the more important it becomes to evaluate them carefully. We must determine whether the information is accurate, whether the recommendation fits the situation, and whether the proposed action moves us toward our purpose.
We also need to consider who could benefit from a decision and who could be harmed by it. Something should not be done simply because technology has made it possible.
The question is no longer whether AI can produce an answer. The question is whether we have the wisdom to know what to do with it.
Discernment turns possibility into direction.
3. Communication: Make the Idea Clear
A powerful idea cannot create change until someone clarifies what matters, simplifies what is complex, and amplifies the right message.
AI allows us to produce more words than ever before, but that does not mean people have more time or attention to give us. Most are already overwhelmed by emails, messages, meetings, reports, posts, and presentations.
Adding more content to the noise is not the same as communicating.
Strong communicators help people understand what matters. They clarify their intention, simplify the message without weakening it, and adapt it for the people who need to hear it. They also confirm that the message was understood.
AI can help us organize, draft, and refine our words. Clear communication, however, still requires a human being who understands the purpose, the audience, and the desired outcome.
Clarity will become a competitive advantage in our noisy, AI-generated world.
4. Connection: Give People a Reason to Care
AI can simulate attention and empathy, but only people can build the trust that makes someone feel seen, heard, and valued.
Technology can help us remember a birthday, prepare for a conversation, or respond more quickly. It can recognize patterns in what someone says and even produce words that sound thoughtful and compassionate.
But words alone do not create a relationship.
Connection requires presence. It grows when we listen to understand, relate through shared experiences, and demonstrate that another person matters to us.
People do not want to feel processed by a system. They want to know that someone genuinely cares about them.
We should use AI to remove the clutter surrounding relationships, not to avoid the responsibility of having them.
As AI becomes more common, authentic human connection will become easier to recognize and harder to replace.
5. Collaboration: Turn the Idea Into Action
AI can assist every person on a team, but it still takes humans to align around a purpose, act together, and achieve something none of them could accomplish alone.
Equipping each team member with an AI assistant does not automatically make the team stronger. People can use excellent tools and still move in different directions, generate polished plans without building agreement, or work faster without working together.
Collaboration requires people to align around what they are trying to accomplish. They must understand their roles, take action, communicate through challenges, and remain accountable to one another.
AI can support that work by organizing information, identifying patterns, documenting decisions, and automating repetitive processes. But technology cannot create unity for us.
People must create the unity.
The greatest accomplishments will not come from technology working alone. They will come from people who know how to combine human strengths with technological capabilities.
Becoming More Human
These five skills create a natural progression. We use creativity to imagine what could be and discernment to decide what should be.
We use communication to make the idea clear and connection to give people a reason to care. Then we use collaboration to turn the idea into action.
AI can support every part of that process, but it should remain an assistant. It should not become the source of our purpose, judgment, relationships, or responsibility.
The goal is to develop the qualities that allow us to use powerful technology wisely. Instead of asking only what AI can do, we must also ask what kind of people we need to become.
While AI may help us do more, these five skills help us become more.
When we learn to guide AI rather than surrender our judgment to it, we can use its power to pursue the goals we choose—not the ones technology chooses for us.
About CJ Powers
CJ Powers is a filmmaker, author, communication strategist, executive speaking coach, and AI consultant. He helps leaders use AI effectively while strengthening the human skills technology cannot replace.
