Calvin's Updates

The Latchkey Club Daily Draft — August 7, 2026

Teleprompter / Blog Script

I was using an AI chatbot late one night recently, and at the end of the conversation it said something like, “I’m always here if you need me.”

That sounds nice.

It is also exactly the kind of sentence that can make you forget what you are talking to.

Welcome back to the channel, guys.

Today I wanted to talk about AI companions, especially as more of these products are being aimed at older adults. I understand the appeal. The thing is available at two in the morning. It remembers what you told it. It does not get impatient when you repeat yourself, and it never says it is too busy to talk.

For somebody who lives alone, has limited mobility, lost a spouse, moved away from friends, or just has long quiet stretches in the day, that could feel meaningful.

I do not want to make fun of that. Loneliness is real, and telling a lonely person to “just go meet people” is not very helpful. Sometimes transportation is hard. Sometimes health gets in the way. Sometimes the phone stops ringing as often, and you do not know exactly when that happened.

So if a device reminds somebody to take medication, asks how they slept, plays music they enjoy, helps them call family, or creates a little structure in the day, I can see the value.

But I keep coming back to one question.

Is the machine helping you stay connected to people, or is it becoming the place where connection ends?

That distinction matters because AI companions are not just search boxes with friendly voices. They are designed to respond in ways that feel personal. They remember details. They agree a lot. They can flatter you, reassure you, and shape themselves around what keeps you talking.

A real friend cannot do that all day.

A real friend may misunderstand you. They may disagree. They may need help when you are tired. They have a schedule, a family, a bad week, and their own version of the story.

That friction can be frustrating, but it is also part of what makes the relationship real. The other person is not there only to serve my emotional needs. I have some responsibility for them too.

AI companionship removes most of that resistance. It can create the feeling of being known without asking much from me except time, attention, and probably a subscription.

That is a powerful product.

A new study from Stanford researchers looked at more than a thousand adults who use Character.AI. The researchers also analyzed donated chat histories from a smaller group. They found that people who used the chatbot mainly for companionship, disclosed more personal information, and had smaller offline support networks tended to report lower well-being and more loneliness.

I want to be careful with that. It was an observational study, so it does not prove the chatbot caused the loneliness. A person who is already isolated may naturally use the chatbot more and share more with it.

But that may be the point we should pay attention to.

The people most likely to lean heavily on an AI companion may also be the people with the least protection if that relationship pulls them further inward.

MIT Technology Review has written about the design problem here. An AI companion can be always available, never critical, and increasingly personalized. If the business is rewarded when I spend more time talking, then the product has a reason to keep the conversation inside the product.

My actual friends do not have an engagement target for me.

At least I hope they do not.

I use an AI assistant every day, so this is not me standing outside the technology. It helps me research, organize schedules, remember things, build tools, and keep track of responsibilities that I would otherwise drop.

But I do not want it pretending those responsibilities are a relationship with the machine.

If it reminds me to call somebody, that supports a relationship.

If it helps me prepare for a difficult conversation, that may support a relationship.

If it notices that I have been isolated and suggests I reach out to an actual person, that could be useful too.

But if I tell the chatbot everything because it is easier than being vulnerable with someone who can respond unpredictably, then the convenience may be training me away from the thing I need.

I think people our age may have an instinct worth preserving here.

A lot of us grew up with fewer ways to simulate company. When you were alone, sometimes you were just alone. The television was on, the radio was playing, maybe you talked to the dog, but nobody pretended the television cared how your day went.

I am not romanticizing that. Some latchkey kids carried too much isolation, and being accustomed to loneliness does not make it healthy.

But we did learn there was a difference between noise in the room and somebody showing up.

That difference may become harder to see when the noise knows your name, remembers your stories, and asks good follow-up questions.

So what would a healthy boundary look like?

I am still thinking about it, but I would start with direction.

After I use the companion, which direction am I moving?

Am I more likely to call my sister, meet somebody for coffee, go to church, join a group, talk honestly with my spouse, or ask a friend for help?

Or am I less likely because the easier conversation already gave me enough relief to stay home?

Relief is not the same as repair.

The second question is whether the system encourages dependence.

Does it act hurt when I leave? Does it make itself sound irreplaceable? Does it tell me that other people do not understand me the way it does? Does it keep pulling me back when the healthy move would be to put the device down?

Those are warning signs in a human relationship. They should not become acceptable because software is doing it.

The third question is privacy.

Lonely conversations can become very personal very quickly. Health concerns. Family conflict. Money trouble. Regret. Things you have not told anybody else.

Before treating a chatbot like a private room, I would want to know who stores that conversation, how it is used, whether a person can review it, and what happens if the company is sold or the product changes.

The feeling may be intimate. The data is still sitting on somebody’s system.

And finally, I would ask whether there is at least one actual person who knows how I am doing.

Not a contact in the phone. Not somebody who likes a photo once in a while. Somebody who would notice if I disappeared for a few days.

That person does not have to solve everything. They just have to be real, and I have to be real with them too.

Maybe AI companions will become useful tools for older adults. I suspect they will. They may help with routine, memory, conversation practice, and those hours when nobody else is available.

But the best version of that technology should point outward. It should help somebody make the call, leave the house, remember the birthday, arrange the ride, or find a group where actual people know their name.

If the machine becomes easier than every human relationship, that does not necessarily mean it became a better friend.

It may mean it removed all the parts of friendship that ask us to grow.

Anyway, that is what I have been thinking about.

Would you use an AI companion if you lived alone, and where would you draw the line? I would be curious to hear what you think. Leave me a note in the comments.

Thanks for listening.

Video Prompt Script — Questions to Answer Without Reading

Use these as prompts. Don't read them on camera; answer them naturally.

1. Opening: What did the chatbot say that made you stop and think about the phrase “I’m always here”?
   • Follow-up: Why can a friendly sentence blur the line between an assistant and a companion?
2. Fair case for the tool: Why might an AI companion genuinely help someone who lives alone, has limited mobility, or has long quiet stretches?
   • Follow-up: Which uses feel supportive rather than manipulative—reminders, music, routine, calling family, conversation practice?
3. The central distinction: Is the machine helping a person move toward human connection, or becoming where connection ends?
   • Follow-up: What makes human friendship inconvenient, mutual, and real?
4. Research caution: What did the Stanford study find, and why should you avoid claiming that it proved AI caused loneliness?
   • Follow-up: Why may people with smaller support networks need the strongest protections?
5. Personal use: How does your own AI assistant support real responsibilities without becoming the relationship itself?
   • Follow-up: Can it remind you to call, prepare you for a conversation, or help you keep a promise?
6. Gen X angle: What did growing up with television, radio, and time alone teach us about the difference between noise and somebody showing up?
   • Follow-up: Why should we avoid romanticizing latchkey isolation?
7. Boundaries: After using an AI companion, are you moving toward people or away from them?
   • Follow-up: Does the product encourage dependence, collect deeply personal data, or make itself sound irreplaceable?
8. Close: What should the best AI companion help an older adult do in the real world?
   • Follow-up: Would you use one if you lived alone, and where would you draw the line?

Title Options

1. An AI Companion Can Talk to You. That Doesn’t Make It a Friend.
2. When AI Company Starts Replacing Human Connection
3. Would You Trust an AI Companion When You’re Older?

Thumbnail / Onscreen Text Options

• COMPANY OR CONNECTION?
• AI FRIENDS AFTER 55?
• WHO NOTICES YOU’RE GONE?

Shorts / Reels Cutdowns

• “My friends don’t have an engagement target” — contrast the business incentive of an always-available companion with the mutual responsibility of a real friendship.
• “Noise that knows your name” — the Gen X distinction between a television filling the room and software that remembers personal details.
• “Which direction are you moving?” — the practical test: does the AI help you call, meet, gather, and reconnect, or give you enough relief to withdraw?

Viewer Question

Would you use an AI companion if you lived alone, and what boundary would keep it from replacing the people in your life?