Last time: 
five games,
several imaginary
missions, and one
legendary box… of
controller cables.

I promised retro
next. Reality
interrupted the
schedule.

Today: 
Call of Duty:
Modern
Warfare 4.

I like the game itself.
But liking a game
does not mean
defending every
choice made around it.

Its opening cinematic
used real footage from
the 2022 Itaewon
crowd-crush disaster
in Seoul.

Real suffering became
scenery for a fictional
war. That is a different
conversation.

This is beyond my
expertise. Meet
Legally Byte,
our legal expert— and
my all-time favorite
reason to invite one.

I prepared
a briefing.
Was there another
reason?

Matching already?
We are dangerously
compatible.

I synchronized
the presentation. Is
that dangerous?

Only to my
concentration.
So—what
actually happened?

Reports 
describe about three
seconds of archive
footage showing an
emergency
response.

It was
repurposed to
depict civilian 
casualties in the
game’s fictional
conflict.

Families
and civic groups
demanded
its removal.

On October 30,
Activision apologized
and announced
removal in an update
scheduled
for November 12.

As of
November 1,
that update is still
ahead. A removal
promise is not
a completed
removal.

So… settled?

A public
response—not
a verified legal
settlement or
damages
award.

The families
welcomed
removal,
but told the BBC
they had not
received a direct
apology.

But did the people
choosing it know
what they were
looking at?

Not just “real
footage.” That
particular tragedy.

Activision
acknowledged
using real news
coverage and denied
intending to evoke or
trivialize the tragedy.

That does
not establish
who knew its exact
origin, or when. 
The public
reports do not
settle that.

And these
reports describe
reused footage.
They do not establish
that AI made
this sequence.

Then AI is our next
question—
not an accusation
about this case.

Exactly. 
No inventing
evidence to make
the lesson tidier.

Suppose they knew
exactly what the 
footage showed
and used it anyway.
Which laws 
could apply?

Let’s use
South Korean
law as our teaching
example. A real
cross-border case
must also resolve
jurisdiction and
applicable law.

First:
copyright. 
The person who
controls the footage
may have a claim if it
was copied without a
license or legal
exception.

Online is not the
same as free?

A door being
open does not
make the
furniture yours.

Permission 
also has a scope.
A news-use license
might not authorize
fictional
entertainment.

And
permission
from the camera’s
owner does not
automatically clear
the rights of the
people filmed.

Two locks. One
key is not enough.

Fair use
is a balancing
test: purpose,
the original work,
how much was
taken, and
market effects.

“Only three
seconds” is
not a magic shield.
Commercial use
is relevant, but not
an automatic loss
either.

For 
identifiable living
people, privacy
and portrait rights
may matter. 
Personal-data
law may also
apply.

Bereaved
families do
not automatically
inherit every privacy
or portrait right.
Any claim based
on their own
protected interests
needs careful legal
analysis.

Different
people.
Different
rights.
Not one
giant
lawsuit
button.

Civil Act
Article 750
covers unlawful
harm caused
intentionally
or negligently.
Article 751
addresses
non-economic
injury.

Knowing the
context can
matter. But a
claimant still needs
a legally protected
interest, unlawful
conduct, harm,
and a causal link.

“That was
insensitive” and
“That creates legal
liability” are
not identical
findings.

Now the other
possibility: they
genuinely did not
know which
tragedy it came
from.

Does “we didn’t know”
make the problem
disappear?

No
automatic 
escape hatch.
Not knowing the
event does not
by itself clear
copyright.

For 
negligence, a
court could ask:
What warning signs
existed? What checks
were reasonable?
Would those checks
have found it?

Especially
when it
visibly
shows
real
people
needing
emergency
help.

What if someone
feeds a random
clip into AI and
says, “Make it
cinematic”?

AI is not a
rights-washing
machine. Check
the input copying
and processing,
then check the
output 
separately.

Changing
the look does not
guarantee that
protected expression
or identifiable
people have
disappeared.

So is there
one law
saying
“Understand
every AI input”?

Do not assume
a  universal rule
requiring complete
historical research
on every file.

Existing 
copyright, privacy,
and negligence rules
can still matter. AI does
not erase them; the duties
depend on the facts.

Our best-practice
checklist: trace the source
and context; check rights and license
scope; review identifiable
people and sensitivity.

Then check the
transformed output,
keep records, and
give a human the
authority to stop
publication.

If a sensitive clip’s
origin is unclear, pause,
investigate, or  replace it. That is
a workflow recommendation—
not a claim that every step
is mandated.

New tools can borrow
reality. They should
not erase the people
who lived it.

Creative
freedom works better
with context—and
accountability.

Excellent.
Now
may I borrow 
you for dinner?

I… have no protocol
for this invitation.