February 2, 2025 • 42 mins
To learn more, please visit Anja Neubauer's site.
Show Notes:
1:30 background in the law and tech
4:45 overview of global framework “AI and Authorship Redefined: Towards a Global Copyright Framework for Commerce and Human Originality - Exploring Ownership, Infringement, Moral Rights, and Human Originality”
12:35 question of authorship
19:35 moral rights
26:30 viability of global copyright
28:20 issues of injustice raised by AI
34:30 German study about developers’ infringement
38:00 public’s choice to lock up data
39:00 The Congress
Please share your comments and/or questions at stephanie@warfareofartandlaw.com
Music by Toulme.
To hear more episodes, please visit Warfare of Art and Law podcast's website.
To leave questions or comments about this or other episodes of the podcast and/or for information about joining the 2ND Saturday discussion on art, culture and justice, please message me at stephanie@warfareofartandlaw.com.
Thanks so much for listening!
© Stephanie Drawdy [2025]
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Speaker 2 (00:28):
I'm fascinated by the evolving implications of tech
on copyright law and Anya sharesabout her research degree
project for her PhD, entitled AIand Authorship Redefined
towards a global copyrightframework for commerce and human
originality, exploringownership infringement, moral
(00:50):
rights and human originality.
Anja Neubauer, welcome toWarfare of Art and Law.
Thank you so much for beinghere.
Speaker 1 (01:02):
Thank you so much for having me here.
Speaker 2 (01:04):
Thank you you so much for being here.
Thank you so much for having mehere.
Thank you, would you give abackground of your history and
your work and what brought youinto this niche area?
Speaker 1 (01:14):
Yes, so I'm a lawyer and I have 20 years plus
experience in copyright law inpractice, which I don't do
anymore, but I'm an expert, soin copyright law and technical
law and when?
So, of course, it was alwaysclear that we have artificial
(01:37):
intelligence, and I'm I'm asci-fi geek, and so it was
obvious that we will have atechnology someday, someday in
far future or nearer future,that would turn everything on
the head, that would makeeverything possible.
We know from the 80s movies,but artificial intelligence was
(02:03):
something that was still faraway and that we had in our
email programs for correctgrammar if you're writing an
email or so.
And then, two years ago, whenOpenAI went public, I saw that
there was a real change.
(02:24):
I saw that there was a realchange, so my head started
exploding of ideas what theconsequences might be then or
now for every branch affected.
And so this is why I firstlythought, okay, I'll do my PhD.
(02:46):
But the deeper I went into thisrabbit hole, so to say, the more
this idea became.
It is necessary that I do this,because there were so many
papers that only said we should,would, could, but nobody dared
to say, okay, this is my draft,let's discuss it, and so I will
(03:13):
make my draft, and I know thereare a lot of mistakes in it and
a lot of things it cannot solveas problems as we see them now,
but at least it's a try.
Afterwards everybody can killme and hate me and and
whatsoever, but at least we haveto try, because I think if we
(03:37):
don't try, we will losecreativity not only job losses,
of course not.
It's.
It's like I think the worldwill always find its way of
living, so nothing is ever lost.
It's.
Imagine the times in the 80s,vhs.
Everybody was in in.
(04:00):
We called it videotechen, sostores where you could rent
those tapes.
And now, if you ask a kidnowadays, do you remember what?
So, but those people are stillalive and they have jobs and
they went on, they developed.
(04:21):
So I'm quite optimistic and Ilove AI.
That's the other hand side.
So yeah, so I'm I'm just um atechnic fan, a sci-fi fan, and
so yeah the, the framework thatyou mentioned.
Speaker 2 (04:42):
Uh, there are 11 parts to it, I I believe.
Would you kind of give anoverview of the framework that
you have developed and just kindof walk us through the overview
of the points that you'retrying to make in your framework
?
Speaker 1 (05:01):
Yes, I will try to do that in just one sentence each.
So, um, there is something likea structure.
These are not articles like inthe law, but just suggestions
what should be done, becauseit's impossible for me alone to
arrange an article, like alegislature would do that, of
(05:25):
course.
So the the first article saysthat we should change the
definition of work, that workshould be defined as an original
creative expression, of course,but that it should be an
integrated thing of semanticcontent, so the ideas, the
(05:52):
meaning and emotional orintellectual expressions, and
the syntactic content, so thestructure, the form, the
technical composition.
And this because it's relatedto the Berne Convention, if you
take this as the treaty thateverybody agreed on and is
confident with.
So this part has to be enlarged, it has to be more than just
(06:20):
this one type of definition.
It has to be enlarged thatthere is also a semantic way.
We have something similar inGerman law, so in the
legislation not it's in the caselaw, so to say, it's called the
principle of the small cointhat at least there must be a
(06:43):
small intent of originality ofsomething that you meant that is
involved in the creativeprocess.
So, and maybe we should enlargethe work definition.
The second is that we have todefine authorship, definitely.
(07:04):
The third article.
There should be requirementsfor licensing agreements, not
only for the AI companies butalso for those who are
recreators of it.
So we also need payment models,licensing models, and we need a
(07:33):
protection.
So that would be article 5 formoral rights that the artists or
the creator is mentioned insome way.
I don't know if the technologywill ever be able to do that,
but I'm quite optimistic because, um, now we have something like
(07:55):
watermarks.
That might be helpful if youcould be, if you could have
something like like watermarking.
Um, yeah, moral rights, so tosay.
I can't even express it whatI'm meaning, just in written
form.
It's easier for me because I'mthe non-native speaker.
I'm sorry for that.
And, um, we need definitely an,uh, an article about
(08:21):
transparency in AI development.
That is very important so thateverybody knows what is
happening, what is inside, whatgoes inside, what is stuck in
there, because this is one ofthe most important questions, if
it's not copyright infringement, from the very second,
(08:41):
something is developed by AI.
Second, something is developedby AI.
A long time myself, I thoughtthat the things that are created
by AI is not a possiblecopyright infringement because
it creates something new.
It's not like in a database andit's flowing out of the vector
arrangement that just imaginesand, um, yeah, reconsiders a new
(09:10):
work.
So this is where I struggled alot, but now it's obvious that
partly it really saves andremembers parts of it, and so
there's always the possibilityof the recreation, and so the
(09:31):
memorizing itself is the thingthat makes it a huge copyright
infringement, so to say.
And of course, in Article 8, Isay that we need exceptions for
educational purposes, because Ithink it's very, very helpful,
(09:54):
especially for training for kids, for elderly people, for I
don't know what else.
For every kind of training it'svery helpful because the AI can
(10:15):
exactly measure at what levelyou are and how to excel your
own capabilities.
So there must also be some kindof exception for that.
Of course, the licensing andattribution is mentioned, the
protection of personal rightsafter the death of a person.
(10:39):
I don't know how many cases.
I have one case that I foundthere was a comedian who died
some years ago and I I recreateda program of him with his voice
and with nowadays jokes, likein the style he would have done
(10:59):
them, which I find veryinteresting, but which I find
very interesting, but it's, it'sspooky.
And then we had the same herein Germany some some weeks ago.
There was an artist.
She died almost 20 years agoand she was in a tv series and
now she is back in the dreams ofher grandchildren and playing
(11:27):
her role again.
It's really spooky.
Maybe you know Gladiator OliverReed died during filming and
with clever techniques of CGIthey reproduced him.
So yeah, would would notice,except if you saw the gladiator
(11:49):
movie with the director's uh, uhcut and yeah.
So this is very important alsohave protection in this, this
area, and of course, and that'sthe most important, I think we
need a global alignment.
Speaker 2 (12:11):
Thank you so much.
That's beautifully done and soso many points to follow up on.
Speaker 1 (12:18):
I'm sorry.
Speaker 2 (12:19):
No, no, it's just because you've covered so much
terrain and you have to withsomething like this.
So just a couple of points Iwanted to pull out.
Could you go back and justspeak a bit more about the
authorship and how you see itdiffering across the three
jurisdictions that you focusedon and what you think needs to
be done with that in a globalapproach?
Speaker 1 (12:42):
yeah, authorship is different.
In the us it says it's, ofcourse, is human, but it's also
possible that computer madeworks are protected so and
(13:04):
that's something that'simpossible in.
Speaker 2 (13:06):
Germany.
Speaker 1 (13:07):
We have this strict authorship that it has to be a
human being, a human intent andtheir menschliche geist.
That's behind it, and in the us, the copyright office is really
struggling with it and making alot of very clever um
(13:32):
suggestions, I think becausethey say we have to make it
different.
On the one hand side, we wedon't want to forbid everything,
because it's also veryimportant that we have AI for
commercial reasons, and so theywant to have a balance between
(13:54):
the, the human author, and theAI co-author, and so there were
several cases like Daya of theDawn and so on, and they tried
to keep a balance somehow.
So in Germany we have noregulation at all, so it sticks
(14:18):
with the human authorship andnothing else is possible.
In my opinion, it should besomething like the suggestions
the copyright office promptingalone can't be as this spark
that is necessary to be reallycreative.
(14:41):
So that was the idea to say youneed to have a semantic and a
syntactic part of a work that iscreated in order to be
protected.
Speaker 2 (14:59):
With that kind of thought, would it in your view
be enough if, say, a drawing bythe human hand is put into the
AI tool and then prompts builton top of that, or a photograph
by human hand with a camera putinto an AI tool?
(15:22):
Would those be sufficient tothen prompt under the model that
you're kind of proposing?
Speaker 1 (15:29):
Possibly, yes, but not the other way around.
But not the other way around.
If I'm just prompting, give mea field of of of red roses or so
, and I get this, and I'm justchanging the prompt a little bit
.
It's not enough creative workof me to make this copyrightable
(15:51):
.
But maybe if you, if you take aphotograph and, for example,
say, uh, make these fields ofroses that I made the photograph
of, make it vivid, and that theroses do something, then I
could think about some kind ofco-production that will be
(16:13):
copyrightable well, because thething is, I mean and this is I
mean I've just been thinkingabout this, the photograph as an
example, because it's even morecompelling than even the
drawing, just knowing thatphotographs, historically, were
challenged, at least in the US.
Speaker 2 (16:30):
I'm not sure how it, what the history of it was in
Germany, If you know, pleaseshare, but the idea that you
know there was a challenge.
Can photographs becopyrightable?
And, yes, they finally decidethey can be.
And now it seems like if aphoto can be copyrightable and
(16:50):
you're putting it into an AItool, it's already copyrightable
.
So then, if you're doingadditional work on it, it seems
like almost a given that if itstarts with a base of your photo
or your drawing and then youprompt with it, it's already
been copyrightable going intothe AI tool.
(17:10):
So how could it magically beuncopyrightable when it comes
out?
Speaker 1 (17:17):
Yes, but you mentioned something very
important it's your copyright,it's your photograph.
It's something different.
And if I say, take what youhave in your pool of vectors and
create something for me andthen try to make arrangements,
that's something completelydifferent.
(17:38):
If you go and arrange,rearrange your photograph, so
why shouldn't it becopyrightable?
Because if you also can makecollages, for example, why not?
Where it becomes very tricky iswhen it is an homage, a pastiche
.
Then it's very, very, very,very difficult.
(18:01):
So a pastiche always.
So by german law, a pastichealways needs this, this human
factor.
There was a possibility, Ithink half a year ago it was
(18:26):
Kraftwerk song again, and theyare fighting for centuries in
the meanwhile.
And now the European court hasto decide if this tiny part of
music in the intro, if that iscopyrightable, if it's a
pastiche, what they use or not.
So the opponent said yes, ofcourse it was a pastiche, and
(18:48):
the other said no, you took itright away.
It's not a pastiche Taking itinto your song.
It can never be a pastichetaking it into your song.
It can never be a pastichebecause.
So this is um now examined bythe european court and it would
be pretty interesting to findout, because from this we could
take some analogies to put intoour other ideas about how to
(19:15):
cope with AI in this particulararea.
So it's a lot of case law Ithink we are facing in the near
future.
Yeah.
Speaker 2 (19:29):
And then the moral rights.
I thought that was so importantthat you put that in there.
Could you kind of speak to that, and what is the current state
of affairs for moral rights inGermany?
Speaker 1 (19:39):
rights, I thought that was so important that you
put that in there.
Could you kind of speak to that?
And what is the current stateof affairs for moral rights in
Germany?
Ah, that is something that Ifound out was very interesting.
I think we have the strongestmoral rights protection from all
legislations.
So in America you only have VARA, and this only protects some
sort of physical things, andthen only in a certain amount
(20:08):
where the name is protected.
So that's weird.
It's really weird.
You don't have it.
You are not obligated to putyour name anywhere.
Instead, in Germany you don'tneed to write a C for copyright
under a picture.
Everybody knows you mustn'tsteal that.
(20:30):
So you have the copyright evenwithout it.
So you have the copyright evenwithout it.
And it might be a tradedisturbance if you put it there.
It might irritate the otherbusinesses around, so you
(20:52):
shouldn't do that.
So in Germany the author'srights are very strong and, to
give you an example, by case lawwas decided.
There was a photographer whosent a cd with his photos to a
magazine and said well, I'm areporter, I make these
(21:14):
photographs.
If you're interested, call me.
This is just an example.
A few weeks later he went to ashop and bought this magazine
and he opened it and saw all hisphotos used there and he wasn't
mentioned.
And then, of course, he suedthe magazine, of course.
(21:36):
And then he said well, the mostimportant thing is, it's not
about the salary, the licenses Iwould have gotten if they would
have bought it from me, but themost disturbing thing is that
they didn't put my name underthe pictures, because now nobody
knows that I'm the greatphotographer.
(21:58):
And then the judges said okay,you're completely right, we must
give you that you could havesold more photographs, but we
don't know how many.
And so let's decide.
You get a so-calledverletzerzuschlag.
So you were disturbed in yourrights and therefore you get the
(22:25):
double fee.
So if the license would be 100euros, then you would get, just
for not mentioning your name,200.
Speaker 2 (22:39):
It's a so-called per photo or just for the whole
incident?
Speaker 1 (22:46):
No, per license, so to say, and this was something
that was used by free onlinepicture databases Very
interesting.
So they said these photographsare for free.
And then the fine print on thehomepage it was written but you
have to mention the name of thephotographer.
(23:07):
So all the people oh, freephotographs.
Put it into the homepages andthey searched for them.
It into the home pages and theysearched for them, and so they
were sued because of, of course,they were allowed to take the
photograph, but they didn'tmention the author, and so they
all had to pay the money justfor not mentioning.
(23:30):
So you see, it's, it's mean andthese were non-quality photos,
and so kleine münze um.
Of course, you can only have aphotograph that's only from a
daisy that you can take.
So it's, it's not really art,we are talking about just stock
(23:50):
pictures, and it was veryfrustrating for lawyers because
of this somehow unfair suingcampaigns that were started, and
later on the fees were reduced,so this was stopped.
But it gives you an idea howcopyright can also be misused in
(24:17):
Germany.
So to stress is what I have tostress is that the moral rights
in Germany are very strict, thatthey really assure your income,
and I also think in some wayit's really necessary that you
(24:43):
as an author, as an artist, youneed that.
So there are only a few caseswhere you really know who the
author was, for example, eminem.
If you hear a song from him,his voice it's so unique, you
immediately know it was him.
But what if AI covered it?
(25:05):
So there we go.
On the one hand side, it's notpossible to have moral rights so
to name the author in a voiceor in a song.
But on the other hand side,everybody knows.
But now we have AI and it'spossible.
What was unthinkable before youreally can cover it.
(25:26):
There was something fromTerminator 2 when the Terminator
covered the voice of the motherof John Connor.
And now we're at that stage andnobody knows how to to handle
this, and this is something thatis.
I only can hope that we havethe technological possibilities
(25:53):
to fix that, that we havesomething at least like
watermarks, to prove thatsomething is original or not, or
what is created or not and thenjust the challenge.
Speaker 2 (26:09):
I mean all the things , all the details and the nuance
of concerns on how, each caseby case, the facts will really
really dictate which waysomething should go, and the
idea that you could have aglobal understanding and
agreement on this just seems.
(26:31):
Some people describe it as verydoable and then, when you get
into the weeds of it, you canstart to feel like it's not
possible.
What would you say to that?
Speaker 1 (26:45):
um, I I.
I know that many people thinkthat it is delusional to say we
will have a global copyright totry, so that seems like a very
optimistic attitude and, yeah, Ithink it's.
(27:12):
It's okay if people laugh aboutit, but I think about all the
original works that need to beprotected in some way.
At least we should try.
I don't know if it works out.
I can't tell you because thetechnological development is so
(27:33):
fast at the moment.
Maybe in two years we have adevelopment that we don't even
discuss that anymore, that weonly laugh about it, because
then it's perfectly clear thatthe originator of any work can
be connected and identifiedimmediately with the wink of an
(27:58):
eye and we don't need to takecare about it.
Speaker 2 (28:03):
I don't know, but at least we should try and think
about it and so the entireconversation to me seems like it
has, uh, concerning points thatwhere you could see there is
injustice occurring or thatcould continue to occur, and so
(28:25):
are there certain points on thatthat you would like to raise
yeah, in, just so.
Speaker 1 (28:35):
The most important thing is that the fair
distribution of rights, therecognition and the
accountability in creativeendeavors and balancing
innovation with human dignity Ithink that's what I would define
(28:58):
justice in this area, and thisis very important to me and, I
think, of course, for all theartists.
So a lot of artists are reallydevastated.
They say how can I make moneywith my work anymore if it's so
(29:18):
inflational?
So you can ask AI, and it doesthe job better.
There were IT experts, aiexperts that said, well, let's
find out.
If AI is better than the humanbeing, then AI would take over.
I don't know, I don't know, butI think, yeah, that's something
(29:50):
that I think we should try toregulate, even if we don't have
the hope to finally change it inthe last second.
So, because AI is there and itwill never go away and there
might be a real chance that, forexample, hollywood is over.
(30:13):
This is what I thought when,last year, zora was introduced
to the public and we see it Allthe directors are joining boards
of ai companies.
David cameron, I read somemonth ago, was joining some
board, maybe just for specialeffects, maybe hopefully, but I
(30:39):
I don't know if he, if he.
But I think something likeoriginality will always be um
the most important thing.
I cannot imagine having, forexample, a house with use huge
pictures.
There are only LCD monitors orso.
(31:01):
These must be real-lifepaintings, and I think that
original paintings, for example,will still be the most precious
thing.
This is something art orartificial intelligence can't
cover in that way.
(31:22):
Of course, there was already anAI that made pictures.
I read that but I think it'ssomething completely different
if it's human painted and youcan ask the artists what was
your imagination?
What did you feel?
What did you feel, what did youthink?
And this together is moreprecious than anything
(31:51):
artificially generated couldever be.
Speaker 2 (31:59):
And that really does hone in on the idea that, in our
current era, even thedefinition of art is becoming
even more of a charged topic,based on what you were just
describing so if you're apainter, I think this is art
that will never die, like beinga talented singer.
Speaker 1 (32:32):
I think there will be ways.
Speaker 2 (32:43):
I don't know why, but I think it will take its time
and it will make its way.
Well, I completely agree that aglobal approach is very needed,
considering we've just lookedat in your PhD, you've just
looked at the threejurisdictions, but there are so
many others that have such awide array of approaches and it
(33:05):
does kind of leave peoplewondering with our in an era
where there is thejurisdictional lines are blurred
so much in the way we exchangeinformation, and how can there
not be a global approach to thisand it will hamper us going
(33:25):
forward if we don't have one.
I I imagine that was yourinspiring thought when you began
this phd thesis yes, takingeverything from everywhere, but
I had to shrink it down.
Speaker 1 (33:39):
The field is way too too huge.
But if you take the best fromthe us law, from the british law
and from our law and try tomake connection, a mixture of it
, taking just the best of it,and then let's have a discussion
, so not only would, could,should, just do it.
(34:00):
So this is more or less thepractical side in me.
That's shouting come on, weneed to do something now.
But I think anyway, it's, it'sreally late, I'm way too late.
I'm I'm way too late with that.
I should.
You need to hurry with thelegislation and also with the
(34:20):
regulation of AI itself.
So this is something thatnobody has to talk about.
So there was a study in Germanythat came out in July.
It also said that OpenAI andall the other companies are huge
(34:44):
copyright infringers becausethe AI is memorizing things and
so therefore, this ispractically always a copyright
infringement, and also thecollection of data by European
legislation.
They say, well, it's allowedfor AI.
No, it's not comparable,because the AI does not only
(35:09):
take the data if it crawlsthrough the internet, it divides
, it takes the syntax and alsothe semantic part of the data,
and so there's not only puredata.
Data would be allowed, but notwith the semantic part of it, so
this is also not allowed.
(35:30):
So, in the end, when, when you,when I, when we use openai,
gemini, whatever they are called, we're using copyrighted
material, or at least parts ofit.
So we are copyright infringers.
You have to say that.
Yes, we are copyrightinfringers.
Speaker 2 (35:53):
Unless the data was ethically taken.
Speaker 1 (35:58):
Unless, unless.
But do we know how did they getthe data?
The idea was in the verybeginning we make open ai open
source.
Everybody could see it, developit, make it better, and then it
was closed and now it's notopen source anymore and now it's
(36:19):
not profitable.
It makes no profits yet.
No, although they rose theprice.
It's now 200 euros if you wantto have the pro version with
Sora in it.
I didn't test Sora yet, but theChinese it's called Tencent.
(36:40):
Just last week they started onevideo or so.
It's open source.
It's mind blowing.
It's so real you can't trust itanymore.
You can't trust your eyes.
It's so real, you see it andit's um.
(37:06):
It has got um consistency incharacters, so they don't change
, and so now it's reallypossible that you can fake
everything flawless.
It's not like taking thepicture of a very popular person
into a video and making somestuff to make it vivid.
(37:29):
You, you actually can copy it.
And I asked myself where didthey get the data from?
Of course, tiktok.
Everybody's uploading its datalike videos and talking to it
and showing the world into thisdatabase.
If they take TikTok, they haveall the information they need.
(37:49):
So a colleague of mine saidlast week.
The only thing we can do is tolock up our data, everybody's
data.
It's the only possibility,because without new and fresh
data, ai will die someday.
And I said, okay, nice idea,but we're living from data,
(38:11):
we're exchanging data all thetime, we are feeding AI with our
data and want suggestions, andso it's a give and take.
I don't think we can stop it.
So, you see, everybody'sthinking about how to stop it on
the one hand side.
On the other hand side, thereare grown-up kids like me that
(38:33):
are watching this and go like,ah, this is outstanding, how
does this work?
Um, but I don't know where it'sending up.
So have you seen, um, the moviethe congress?
So the the new york times cameup with when, when sylva was
(38:55):
published not for public, butwhen they said, okay, we have it
and we will test it and thenwe'll come, come up with it.
Uh, the new york times had anarticle that was called um, what
was it like?
Uh did?
Did the movie the congress seethe future 10 years ago?
(39:15):
Something like that?
That was the title, and I wascurious because they said it was
a movie.
Um, robin wright was playingherself and she was an artist,
that is, that that went olderand she was asked to copy
(39:35):
herself into an ai, and then shewas.
20 years later she was invitedto a congress and then they
wanted to convince her that shewould also give AI her smell and
her body chemicals, because itwas would be the news trend that
people would swallow a pill,that they would be her and feel
(40:01):
like her.
It's a completely crazy movie,but in the end it turns out the
real world is only forlow-budget earners who are just
taking pills in order to imagineto be in the real world where
they feel fine but everything isin decay and rubble.
(40:26):
And it's very thoughtful, Ithink, because it shows how
things that are possible todaymight turn out in the far future
, and so I think all thesecreepy experiences in such
movies are helping to clear yourmind, to see what we have to
(40:52):
demand to keep our families, ourfuture of our children safe in
future our children safe infuture.
Speaker 2 (41:04):
There will be links in the show notes to learn more
about Anya Neubauer and her work.
If you were intrigued by thispodcast, it'd be much
appreciated if you would leave arating or review and tag
Warfare of Art and Law podcast.
Until next time, this isStephanie Drotty bringing you
Warfare of Art and Law.
Thank you so much for listeningand remember injustice anywhere
(41:26):
is a threat to justiceeverywhere.
I'm fascinated by the evolving implications of tech
on copyright law and Anya sharesabout her research degree
project for her PhD, entitled AIand Authorship Redefined
towards a global copyrightframework for commerce and human
originality, exploringownership infringement, moral
(00:50):
rights and human originality.
Anja Neubauer, welcome toWarfare of Art and Law.
Thank you so much for beinghere.
Speaker 1 (01:02):
Thank you so much for having me here.
Speaker 2 (01:04):
Thank you you so much for being here.
Thank you so much for having mehere.
Thank you, would you give abackground of your history and
your work and what brought youinto this niche area?
Speaker 1 (01:14):
Yes, so I'm a lawyer and I have 20 years plus
experience in copyright law inpractice, which I don't do
anymore, but I'm an expert, soin copyright law and technical
law and when?
So, of course, it was alwaysclear that we have artificial
(01:37):
intelligence, and I'm I'm asci-fi geek, and so it was
obvious that we will have atechnology someday, someday in
far future or nearer future,that would turn everything on
the head, that would makeeverything possible.
We know from the 80s movies,but artificial intelligence was
(02:03):
something that was still faraway and that we had in our
email programs for correctgrammar if you're writing an
email or so.
And then, two years ago, whenOpenAI went public, I saw that
there was a real change.
(02:24):
I saw that there was a realchange, so my head started
exploding of ideas what theconsequences might be then or
now for every branch affected.
And so this is why I firstlythought, okay, I'll do my PhD.
(02:46):
But the deeper I went into thisrabbit hole, so to say, the more
this idea became.
It is necessary that I do this,because there were so many
papers that only said we should,would, could, but nobody dared
to say, okay, this is my draft,let's discuss it, and so I will
(03:13):
make my draft, and I know thereare a lot of mistakes in it and
a lot of things it cannot solveas problems as we see them now,
but at least it's a try.
Afterwards everybody can killme and hate me and and
whatsoever, but at least we haveto try, because I think if we
(03:37):
don't try, we will losecreativity not only job losses,
of course not.
It's.
It's like I think the worldwill always find its way of
living, so nothing is ever lost.
It's.
Imagine the times in the 80s,vhs.
Everybody was in in.
(04:00):
We called it videotechen, sostores where you could rent
those tapes.
And now, if you ask a kidnowadays, do you remember what?
So, but those people are stillalive and they have jobs and
they went on, they developed.
(04:21):
So I'm quite optimistic and Ilove AI.
That's the other hand side.
So yeah, so I'm I'm just um atechnic fan, a sci-fi fan, and
so yeah the, the framework thatyou mentioned.
Speaker 2 (04:42):
Uh, there are 11 parts to it, I I believe.
Would you kind of give anoverview of the framework that
you have developed and just kindof walk us through the overview
of the points that you'retrying to make in your framework
?
Speaker 1 (05:01):
Yes, I will try to do that in just one sentence each.
So, um, there is something likea structure.
These are not articles like inthe law, but just suggestions
what should be done, becauseit's impossible for me alone to
arrange an article, like alegislature would do that, of
(05:25):
course.
So the the first article saysthat we should change the
definition of work, that workshould be defined as an original
creative expression, of course,but that it should be an
integrated thing of semanticcontent, so the ideas, the
(05:52):
meaning and emotional orintellectual expressions, and
the syntactic content, so thestructure, the form, the
technical composition.
And this because it's relatedto the Berne Convention, if you
take this as the treaty thateverybody agreed on and is
confident with.
So this part has to be enlarged, it has to be more than just
(06:20):
this one type of definition.
It has to be enlarged thatthere is also a semantic way.
We have something similar inGerman law, so in the
legislation not it's in the caselaw, so to say, it's called the
principle of the small cointhat at least there must be a
(06:43):
small intent of originality ofsomething that you meant that is
involved in the creativeprocess.
So, and maybe we should enlargethe work definition.
The second is that we have todefine authorship, definitely.
(07:04):
The third article.
There should be requirementsfor licensing agreements, not
only for the AI companies butalso for those who are
recreators of it.
So we also need payment models,licensing models, and we need a
(07:33):
protection.
So that would be article 5 formoral rights that the artists or
the creator is mentioned insome way.
I don't know if the technologywill ever be able to do that,
but I'm quite optimistic because, um, now we have something like
(07:55):
watermarks.
That might be helpful if youcould be, if you could have
something like like watermarking.
Um, yeah, moral rights, so tosay.
I can't even express it whatI'm meaning, just in written
form.
It's easier for me because I'mthe non-native speaker.
I'm sorry for that.
And, um, we need definitely an,uh, an article about
(08:21):
transparency in AI development.
That is very important so thateverybody knows what is
happening, what is inside, whatgoes inside, what is stuck in
there, because this is one ofthe most important questions, if
it's not copyright infringement, from the very second,
(08:41):
something is developed by AI.
Second, something is developedby AI.
A long time myself, I thoughtthat the things that are created
by AI is not a possiblecopyright infringement because
it creates something new.
It's not like in a database andit's flowing out of the vector
arrangement that just imaginesand, um, yeah, reconsiders a new
(09:10):
work.
So this is where I struggled alot, but now it's obvious that
partly it really saves andremembers parts of it, and so
there's always the possibilityof the recreation, and so the
(09:31):
memorizing itself is the thingthat makes it a huge copyright
infringement, so to say.
And of course, in Article 8, Isay that we need exceptions for
educational purposes, because Ithink it's very, very helpful,
(09:54):
especially for training for kids, for elderly people, for I
don't know what else.
For every kind of training it'svery helpful because the AI can
(10:15):
exactly measure at what levelyou are and how to excel your
own capabilities.
So there must also be some kindof exception for that.
Of course, the licensing andattribution is mentioned, the
protection of personal rightsafter the death of a person.
(10:39):
I don't know how many cases.
I have one case that I foundthere was a comedian who died
some years ago and I I recreateda program of him with his voice
and with nowadays jokes, likein the style he would have done
(10:59):
them, which I find veryinteresting, but which I find
very interesting, but it's, it'sspooky.
And then we had the same herein Germany some some weeks ago.
There was an artist.
She died almost 20 years agoand she was in a tv series and
now she is back in the dreams ofher grandchildren and playing
(11:27):
her role again.
It's really spooky.
Maybe you know Gladiator OliverReed died during filming and
with clever techniques of CGIthey reproduced him.
So yeah, would would notice,except if you saw the gladiator
(11:49):
movie with the director's uh, uhcut and yeah.
So this is very important alsohave protection in this, this
area, and of course, and that'sthe most important, I think we
need a global alignment.
Speaker 2 (12:11):
Thank you so much.
That's beautifully done and soso many points to follow up on.
Speaker 1 (12:18):
I'm sorry.
Speaker 2 (12:19):
No, no, it's just because you've covered so much
terrain and you have to withsomething like this.
So just a couple of points Iwanted to pull out.
Could you go back and justspeak a bit more about the
authorship and how you see itdiffering across the three
jurisdictions that you focusedon and what you think needs to
be done with that in a globalapproach?
Speaker 1 (12:42):
yeah, authorship is different.
In the us it says it's, ofcourse, is human, but it's also
possible that computer madeworks are protected so and
(13:04):
that's something that'simpossible in.
Speaker 2 (13:06):
Germany.
Speaker 1 (13:07):
We have this strict authorship that it has to be a
human being, a human intent andtheir menschliche geist.
That's behind it, and in the us, the copyright office is really
struggling with it and making alot of very clever um
(13:32):
suggestions, I think becausethey say we have to make it
different.
On the one hand side, we wedon't want to forbid everything,
because it's also veryimportant that we have AI for
commercial reasons, and so theywant to have a balance between
(13:54):
the, the human author, and theAI co-author, and so there were
several cases like Daya of theDawn and so on, and they tried
to keep a balance somehow.
So in Germany we have noregulation at all, so it sticks
(14:18):
with the human authorship andnothing else is possible.
In my opinion, it should besomething like the suggestions
the copyright office promptingalone can't be as this spark
that is necessary to be reallycreative.
(14:41):
So that was the idea to say youneed to have a semantic and a
syntactic part of a work that iscreated in order to be
protected.
Speaker 2 (14:59):
With that kind of thought, would it in your view
be enough if, say, a drawing bythe human hand is put into the
AI tool and then prompts builton top of that, or a photograph
by human hand with a camera putinto an AI tool?
(15:22):
Would those be sufficient tothen prompt under the model that
you're kind of proposing?
Speaker 1 (15:29):
Possibly, yes, but not the other way around.
But not the other way around.
If I'm just prompting, give mea field of of of red roses or so
, and I get this, and I'm justchanging the prompt a little bit
.
It's not enough creative workof me to make this copyrightable
(15:51):
.
But maybe if you, if you take aphotograph and, for example,
say, uh, make these fields ofroses that I made the photograph
of, make it vivid, and that theroses do something, then I
could think about some kind ofco-production that will be
(16:13):
copyrightable well, because thething is, I mean and this is I
mean I've just been thinkingabout this, the photograph as an
example, because it's even morecompelling than even the
drawing, just knowing thatphotographs, historically, were
challenged, at least in the US.
Speaker 2 (16:30):
I'm not sure how it, what the history of it was in
Germany, If you know, pleaseshare, but the idea that you
know there was a challenge.
Can photographs becopyrightable?
And, yes, they finally decidethey can be.
And now it seems like if aphoto can be copyrightable and
(16:50):
you're putting it into an AItool, it's already copyrightable
.
So then, if you're doingadditional work on it, it seems
like almost a given that if itstarts with a base of your photo
or your drawing and then youprompt with it, it's already
been copyrightable going intothe AI tool.
(17:10):
So how could it magically beuncopyrightable when it comes
out?
Speaker 1 (17:17):
Yes, but you mentioned something very
important it's your copyright,it's your photograph.
It's something different.
And if I say, take what youhave in your pool of vectors and
create something for me andthen try to make arrangements,
that's something completelydifferent.
(17:38):
If you go and arrange,rearrange your photograph, so
why shouldn't it becopyrightable?
Because if you also can makecollages, for example, why not?
Where it becomes very tricky iswhen it is an homage, a pastiche
.
Then it's very, very, very,very difficult.
(18:01):
So a pastiche always.
So by german law, a pastichealways needs this, this human
factor.
There was a possibility, Ithink half a year ago it was
(18:26):
Kraftwerk song again, and theyare fighting for centuries in
the meanwhile.
And now the European court hasto decide if this tiny part of
music in the intro, if that iscopyrightable, if it's a
pastiche, what they use or not.
So the opponent said yes, ofcourse it was a pastiche, and
(18:48):
the other said no, you took itright away.
It's not a pastiche Taking itinto your song.
It can never be a pastichetaking it into your song.
It can never be a pastichebecause.
So this is um now examined bythe european court and it would
be pretty interesting to findout, because from this we could
take some analogies to put intoour other ideas about how to
(19:15):
cope with AI in this particulararea.
So it's a lot of case law Ithink we are facing in the near
future.
Yeah.
Speaker 2 (19:29):
And then the moral rights.
I thought that was so importantthat you put that in there.
Could you kind of speak to that, and what is the current state
of affairs for moral rights inGermany?
Speaker 1 (19:39):
rights, I thought that was so important that you
put that in there.
Could you kind of speak to that?
And what is the current stateof affairs for moral rights in
Germany?
Ah, that is something that Ifound out was very interesting.
I think we have the strongestmoral rights protection from all
legislations.
So in America you only have VARA, and this only protects some
sort of physical things, andthen only in a certain amount
(20:08):
where the name is protected.
So that's weird.
It's really weird.
You don't have it.
You are not obligated to putyour name anywhere.
Instead, in Germany you don'tneed to write a C for copyright
under a picture.
Everybody knows you mustn'tsteal that.
(20:30):
So you have the copyright evenwithout it.
So you have the copyright evenwithout it.
And it might be a tradedisturbance if you put it there.
It might irritate the otherbusinesses around, so you
(20:52):
shouldn't do that.
So in Germany the author'srights are very strong and, to
give you an example, by case lawwas decided.
There was a photographer whosent a cd with his photos to a
magazine and said well, I'm areporter, I make these
(21:14):
photographs.
If you're interested, call me.
This is just an example.
A few weeks later he went to ashop and bought this magazine
and he opened it and saw all hisphotos used there and he wasn't
mentioned.
And then, of course, he suedthe magazine, of course.
(21:36):
And then he said well, the mostimportant thing is, it's not
about the salary, the licenses Iwould have gotten if they would
have bought it from me, but themost disturbing thing is that
they didn't put my name underthe pictures, because now nobody
knows that I'm the greatphotographer.
(21:58):
And then the judges said okay,you're completely right, we must
give you that you could havesold more photographs, but we
don't know how many.
And so let's decide.
You get a so-calledverletzerzuschlag.
So you were disturbed in yourrights and therefore you get the
(22:25):
double fee.
So if the license would be 100euros, then you would get, just
for not mentioning your name,200.
Speaker 2 (22:39):
It's a so-called per photo or just for the whole
incident?
Speaker 1 (22:46):
No, per license, so to say, and this was something
that was used by free onlinepicture databases Very
interesting.
So they said these photographsare for free.
And then the fine print on thehomepage it was written but you
have to mention the name of thephotographer.
(23:07):
So all the people oh, freephotographs.
Put it into the homepages andthey searched for them.
It into the home pages and theysearched for them, and so they
were sued because of, of course,they were allowed to take the
photograph, but they didn'tmention the author, and so they
all had to pay the money justfor not mentioning.
(23:30):
So you see, it's, it's mean andthese were non-quality photos,
and so kleine münze um.
Of course, you can only have aphotograph that's only from a
daisy that you can take.
So it's, it's not really art,we are talking about just stock
(23:50):
pictures, and it was veryfrustrating for lawyers because
of this somehow unfair suingcampaigns that were started, and
later on the fees were reduced,so this was stopped.
But it gives you an idea howcopyright can also be misused in
(24:17):
Germany.
So to stress is what I have tostress is that the moral rights
in Germany are very strict, thatthey really assure your income,
and I also think in some wayit's really necessary that you
(24:43):
as an author, as an artist, youneed that.
So there are only a few caseswhere you really know who the
author was, for example, eminem.
If you hear a song from him,his voice it's so unique, you
immediately know it was him.
But what if AI covered it?
(25:05):
So there we go.
On the one hand side, it's notpossible to have moral rights so
to name the author in a voiceor in a song.
But on the other hand side,everybody knows.
But now we have AI and it'spossible.
What was unthinkable before youreally can cover it.
(25:26):
There was something fromTerminator 2 when the Terminator
covered the voice of the motherof John Connor.
And now we're at that stage andnobody knows how to to handle
this, and this is something thatis.
I only can hope that we havethe technological possibilities
(25:53):
to fix that, that we havesomething at least like
watermarks, to prove thatsomething is original or not, or
what is created or not and thenjust the challenge.
Speaker 2 (26:09):
I mean all the things , all the details and the nuance
of concerns on how, each caseby case, the facts will really
really dictate which waysomething should go, and the
idea that you could have aglobal understanding and
agreement on this just seems.
(26:31):
Some people describe it as verydoable and then, when you get
into the weeds of it, you canstart to feel like it's not
possible.
What would you say to that?
Speaker 1 (26:45):
um, I I.
I know that many people thinkthat it is delusional to say we
will have a global copyright totry, so that seems like a very
optimistic attitude and, yeah, Ithink it's.
(27:12):
It's okay if people laugh aboutit, but I think about all the
original works that need to beprotected in some way.
At least we should try.
I don't know if it works out.
I can't tell you because thetechnological development is so
(27:33):
fast at the moment.
Maybe in two years we have adevelopment that we don't even
discuss that anymore, that weonly laugh about it, because
then it's perfectly clear thatthe originator of any work can
be connected and identifiedimmediately with the wink of an
(27:58):
eye and we don't need to takecare about it.
Speaker 2 (28:03):
I don't know, but at least we should try and think
about it and so the entireconversation to me seems like it
has, uh, concerning points thatwhere you could see there is
injustice occurring or thatcould continue to occur, and so
(28:25):
are there certain points on thatthat you would like to raise
yeah, in, just so.
Speaker 1 (28:35):
The most important thing is that the fair
distribution of rights, therecognition and the
accountability in creativeendeavors and balancing
innovation with human dignity Ithink that's what I would define
(28:58):
justice in this area, and thisis very important to me and, I
think, of course, for all theartists.
So a lot of artists are reallydevastated.
They say how can I make moneywith my work anymore if it's so
(29:18):
inflational?
So you can ask AI, and it doesthe job better.
There were IT experts, aiexperts that said, well, let's
find out.
If AI is better than the humanbeing, then AI would take over.
I don't know, I don't know, butI think, yeah, that's something
(29:50):
that I think we should try toregulate, even if we don't have
the hope to finally change it inthe last second.
So, because AI is there and itwill never go away and there
might be a real chance that, forexample, hollywood is over.
(30:13):
This is what I thought when,last year, zora was introduced
to the public and we see it Allthe directors are joining boards
of ai companies.
David cameron, I read somemonth ago, was joining some
board, maybe just for specialeffects, maybe hopefully, but I
(30:39):
I don't know if he, if he.
But I think something likeoriginality will always be um
the most important thing.
I cannot imagine having, forexample, a house with use huge
pictures.
There are only LCD monitors orso.
(31:01):
These must be real-lifepaintings, and I think that
original paintings, for example,will still be the most precious
thing.
This is something art orartificial intelligence can't
cover in that way.
(31:22):
Of course, there was already anAI that made pictures.
I read that but I think it'ssomething completely different
if it's human painted and youcan ask the artists what was
your imagination?
What did you feel?
What did you feel, what did youthink?
And this together is moreprecious than anything
(31:51):
artificially generated couldever be.
Speaker 2 (31:59):
And that really does hone in on the idea that, in our
current era, even thedefinition of art is becoming
even more of a charged topic,based on what you were just
describing so if you're apainter, I think this is art
that will never die, like beinga talented singer.
Speaker 1 (32:32):
I think there will be ways.
Speaker 2 (32:43):
I don't know why, but I think it will take its time
and it will make its way.
Well, I completely agree that aglobal approach is very needed,
considering we've just lookedat in your PhD, you've just
looked at the threejurisdictions, but there are so
many others that have such awide array of approaches and it
(33:05):
does kind of leave peoplewondering with our in an era
where there is thejurisdictional lines are blurred
so much in the way we exchangeinformation, and how can there
not be a global approach to thisand it will hamper us going
(33:25):
forward if we don't have one.
I I imagine that was yourinspiring thought when you began
this phd thesis yes, takingeverything from everywhere, but
I had to shrink it down.
Speaker 1 (33:39):
The field is way too too huge.
But if you take the best fromthe us law, from the british law
and from our law and try tomake connection, a mixture of it
, taking just the best of it,and then let's have a discussion
, so not only would, could,should, just do it.
(34:00):
So this is more or less thepractical side in me.
That's shouting come on, weneed to do something now.
But I think anyway, it's, it'sreally late, I'm way too late.
I'm I'm way too late with that.
I should.
You need to hurry with thelegislation and also with the
(34:20):
regulation of AI itself.
So this is something thatnobody has to talk about.
So there was a study in Germanythat came out in July.
It also said that OpenAI andall the other companies are huge
(34:44):
copyright infringers becausethe AI is memorizing things and
so therefore, this ispractically always a copyright
infringement, and also thecollection of data by European
legislation.
They say, well, it's allowedfor AI.
No, it's not comparable,because the AI does not only
(35:09):
take the data if it crawlsthrough the internet, it divides
, it takes the syntax and alsothe semantic part of the data,
and so there's not only puredata.
Data would be allowed, but notwith the semantic part of it, so
this is also not allowed.
(35:30):
So, in the end, when, when you,when I, when we use openai,
gemini, whatever they are called, we're using copyrighted
material, or at least parts ofit.
So we are copyright infringers.
You have to say that.
Yes, we are copyrightinfringers.
Speaker 2 (35:53):
Unless the data was ethically taken.
Speaker 1 (35:58):
Unless, unless.
But do we know how did they getthe data?
The idea was in the verybeginning we make open ai open
source.
Everybody could see it, developit, make it better, and then it
was closed and now it's notopen source anymore and now it's
(36:19):
not profitable.
It makes no profits yet.
No, although they rose theprice.
It's now 200 euros if you wantto have the pro version with
Sora in it.
I didn't test Sora yet, but theChinese it's called Tencent.
(36:40):
Just last week they started onevideo or so.
It's open source.
It's mind blowing.
It's so real you can't trust itanymore.
You can't trust your eyes.
It's so real, you see it andit's um.
(37:06):
It has got um consistency incharacters, so they don't change
, and so now it's reallypossible that you can fake
everything flawless.
It's not like taking thepicture of a very popular person
into a video and making somestuff to make it vivid.
(37:29):
You, you actually can copy it.
And I asked myself where didthey get the data from?
Of course, tiktok.
Everybody's uploading its datalike videos and talking to it
and showing the world into thisdatabase.
If they take TikTok, they haveall the information they need.
(37:49):
So a colleague of mine saidlast week.
The only thing we can do is tolock up our data, everybody's
data.
It's the only possibility,because without new and fresh
data, ai will die someday.
And I said, okay, nice idea,but we're living from data,
(38:11):
we're exchanging data all thetime, we are feeding AI with our
data and want suggestions, andso it's a give and take.
I don't think we can stop it.
So, you see, everybody'sthinking about how to stop it on
the one hand side.
On the other hand side, thereare grown-up kids like me that
(38:33):
are watching this and go like,ah, this is outstanding, how
does this work?
Um, but I don't know where it'sending up.
So have you seen, um, the moviethe congress?
So the the new york times cameup with when, when sylva was
(38:55):
published not for public, butwhen they said, okay, we have it
and we will test it and thenwe'll come, come up with it.
Uh, the new york times had anarticle that was called um, what
was it like?
Uh did?
Did the movie the congress seethe future 10 years ago?
(39:15):
Something like that?
That was the title, and I wascurious because they said it was
a movie.
Um, robin wright was playingherself and she was an artist,
that is, that that went olderand she was asked to copy
(39:35):
herself into an ai, and then shewas.
20 years later she was invitedto a congress and then they
wanted to convince her that shewould also give AI her smell and
her body chemicals, because itwas would be the news trend that
people would swallow a pill,that they would be her and feel
(40:01):
like her.
It's a completely crazy movie,but in the end it turns out the
real world is only forlow-budget earners who are just
taking pills in order to imagineto be in the real world where
they feel fine but everything isin decay and rubble.
(40:26):
And it's very thoughtful, Ithink, because it shows how
things that are possible todaymight turn out in the far future
, and so I think all thesecreepy experiences in such
movies are helping to clear yourmind, to see what we have to
(40:52):
demand to keep our families, ourfuture of our children safe in
future our children safe infuture.
Speaker 2 (41:04):
There will be links in the show notes to learn more
about Anya Neubauer and her work.
If you were intrigued by thispodcast, it'd be much
appreciated if you would leave arating or review and tag
Warfare of Art and Law podcast.
Until next time, this isStephanie Drotty bringing you
Warfare of Art and Law.
Thank you so much for listeningand remember injustice anywhere
(41:26):
is a threat to justiceeverywhere.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
The Breakfast Club
The World's Most Dangerous Morning Show, The Breakfast Club, With DJ Envy And Charlamagne Tha God!
The Joe Rogan Experience
The official podcast of comedian Joe Rogan.