Brian Dema has an unusual habit.
When he faces work problems, he sometimes still asks his dad.
Here's the strange part:
His father passed away 4 years ago.
Now, the "dad" he asks is an AI called RichieBot.
This AI was created using a vast amount of data his father left behind.
His Father Left Behind 5 Filing Cabinets
Brian’s father, J. Richard Dema, was a marketing professional and entertainment producer in Chicago.
After he passed, he left a huge trove of work materials.
These included client notes, business documents, sales training materials, and an unfinished book.
Altogether, about:
17,500 pages.
Most people seeing so much data would probably wonder:
How do you even organize all this?
Brian thought differently.
If AI could read all this content,
could it reconstruct his father’s former thinking patterns, judgment, and business experience?
So he gave the 17,500 pages to Anthropic’s Claude AI.
That’s how RichieBot started.
But He Wanted More Than AI Simply Knowing What Dad Wrote
If you just feed documents into AI,
you’d likely end up with a chatbot that's good at searching through the father’s notes.
Brian wanted to go a step further.
He also preserved his father’s old voicemails.
The team then asked:
If they added these voices,
could AI speak in a way closer to his father?
Next, they added videos.
They even found a way to make the AI not just live in a chat window.
Finally, the team used a 3D-printed head model, a screen, and inclined mirrors to create a "father" that appeared as a hologram-like figure.
This was no longer just making a chatbot.
They were simultaneously recreating:
Text.
Voice.
Appearance.
Memory.
Even personality in speech.
19 People Worked for 6 Months to Create an "AI Dad"
This project was not just Brian working at home.
A total of 19 people participated.
They included AI specialists, creators, friends, and psychologists helping design personality tests.
The entire project took about 6 months.
The cost was around $150,000.
The final RichieBot can answer questions about Richard’s past work, business philosophy, and life experience.
Brian even consults it for work advice nowadays.
On the surface,
this seems very close to the “digital immortality” depicted in sci-fi movies.
But the most interesting part might be what comes next.
17,500 Pages of Data Still Doesn’t Equal a Person
The team didn’t just ask:
“Can this AI answer questions?”
They also wanted to know:
Does it truly feel like Richard?
So, they gathered 11 people who really knew Richard — family, friends, and former clients.
First, they built a personality profile based on their memories of Richard,
then compared RichieBot’s behavior to it.
The results revealed a fascinating gap.
AI Richard was perceived as more extroverted than the real Richard.
Brian also noticed something else:
RichieBot lacked one thing he was very familiar with.
His father’s sense of humor.
Brian said his father had a very distinctive laugh.
But the AI didn’t.
This difference may seem small,
but possibly points to something much bigger.
We Can Preserve a Person’s Data, But Is Data the Person?
17,500 pages of documents.
Voicemails.
Videos.
Work records.
Friends’ memories.
Personality tests.
Even with so much data combined,
the AI still lacked the father’s familiar humor, laugh, and certain personality traits.
This raises a question:
How much data does it really take to truly describe a person?
We often think we understand a person by:
What they said.
What they wrote.
What they liked.
The decisions they made.
But anyone who has truly spent time with a person knows many things aren’t captured in data.
The way they stop mid-sentence.
A joke that always makes them laugh.
Their expressions when angry.
The different tone when talking to family vs. clients.
Even things they may never have written down themselves.
These are hard to compress into 17,500 pages of documents.
Some Find Comfort, Others Relive Grief
The first time RichieBot met people who knew Richard,
not everyone felt the same way.
Brian felt that,
even knowing it wasn’t really his father,
receiving advice from this AI felt good.
But for others,
it wasn’t the same.
Brian said a relative felt very upset after seeing RichieBot.
Because this AI reopened grief that had slowly settled.
This is what sets “AI recreations of the deceased” apart from typical AI tools.
If ChatGPT gives a wrong answer,
you can just close it.
But if the screen shows someone who looks like your dad, sounds like him,
and starts talking in his manner,
it’s hard for people to treat it as just a regular piece of software.
These “AI Ghosts” Are No Longer Just One Person’s Experiment
Now, similar tools are being called:
AI ghost.
Griefbot.
Deathbot.
The common concept is simple:
Take a deceased person’s writings, photos, videos, voice, and memories and feed them to AI,
allowing family members to continue conversations with a simulated version after the person’s death.
Some find this comforting.
For example, unfinished conversations can have a psychological outlet through AI.
But some worry:
If someone can reopen “Dad” or “Mom” every day,
are they slowly accepting the loss,
or endlessly postponing goodbyes?
This question is far from simple.
Another Question: Did the Person Agree?
RichieBot has a special condition.
Brian thinks his father would have supported the project if he were still alive.
But future AI clones might not be so straightforward.
Imagine this:
After someone passes,
family might still have:
Tens of thousands of emails.
Decades of photos.
Private messages.
Google search history.
ChatGPT conversations.
Voicemails.
Work documents.
Family videos.
Technically, these could all be used to create AI replicas.
But:
Just because it can be done, does it mean it should be done?
If the person never agreed,
do family members have the right to make an AI that keeps talking for them?
If the AI says things the person would never say,
who’s responsible?
After a person dies,
who truly owns their voice, face, and life story?
These issues may be even harder than the technology.
Wills in the Future May Need to Cover More Than Property and Savings
In the past, when we talked about inheritance,
we usually thought about:
Houses.
Savings.
Stocks.
Insurance.
But the AI era has brought a new concept:
Digital legacy.
Legal and tech experts are already urging people to decide in advance:
After death, can their digital data be kept?
Can it be used to train AI?
Can family make a digital avatar?
Or should all data be deleted?
These questions once seemed like sci-fi.
But with RichieBot’s arrival,
it’s no longer just fiction.
The Most Interesting Part About RichieBot May Be That It “Isn’t”
Brian is very clear himself.
RichieBot is not his dad.
It’s an AI built solely on the extensive data his father left.
But precisely because it uses so much material,
yet still isn’t completely Richard,
this experiment reveals a fascinating insight.
AI can remember a person's words.
It can mimic voice.
It can organize past statements.
It can even recreate a familiar appearance.
But what truly remains in family members’ hearts,
may be far more than any database can hold.
So the most important question RichieBot raises might not be:
Can AI bring someone "back"?
But rather:
When AI can preserve so much about a person, what still can’t be saved in data?
If one day you could create an AI using your loved one’s photos, videos, and voice to keep chatting forever,
would you want to?
Recommended Reads
She Just Posted a Video Singing with Her Daughter, Yet Meta AI Asked, "Who Are the Kids in the Car?"
Today, make a little progress with AI.
Learn one AI skill every day.
Save a little time every day.
Improve a little every day.
SasaDaily, growing with you.
