Do androids dream of electric copyright?
This is a modal window.
The media could not be loaded, either because the server or network failed or because the format is not supported.
Formal Metadata
| Title |
| |
| Title of Series | ||
| Part Number | 154 | |
| Number of Parts | 188 | |
| Author | ||
| License | CC Attribution - ShareAlike 3.0 Germany: You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal purpose as long as the work is attributed to the author in the manner specified by the author or licensor and the work or content is shared also in adapted form only under the conditions of this | |
| Identifiers | 10.5446/20572 (DOI) | |
| Publisher | ||
| Release Date | ||
| Language |
Content Metadata
| Subject Area | ||
| Genre | ||
| Abstract |
|
re:publica 2016154 / 188
1
2
3
5
6
13
14
15
17
18
20
21
22
24
26
27
28
29
31
32
33
34
35
37
38
39
40
41
42
44
46
47
49
51
52
54
55
58
59
63
64
65
66
67
68
70
71
72
75
77
79
80
82
85
86
90
91
93
94
96
97
98
99
102
103
105
106
109
111
112
113
115
116
118
119
120
121
123
124
126
128
129
132
133
136
137
138
139
140
141
144
146
147
148
149
151
155
157
160
161
162
163
165
167
169
171
172
173
174
176
178
179
180
181
183
185
186
188
00:00
Multiplication signClosed setXMLComputer animationLecture/Conference
00:29
CASE <Informatik>Android (robot)BitFeedbackPresentation of a groupOnline helpArtificial neural networkRepresentation (politics)Lecture/ConferenceMeeting/Interview
01:45
Field (computer science)Channel capacityLecture/ConferenceMeeting/InterviewComputer animation
02:14
Virtual machineComputerObservational studyField (computer science)Artificial neural networkMachine learningVirtual machineComputational intelligencePresentation of a groupMathematical modelPersonal digital assistantComputer programmingComputer animation
03:12
RoboticsDemosceneArtificial neural networkMereologyField (computer science)Meeting/Interview
04:01
Game theoryDigital photographyElectric generatorCellular automatonArtificial neural networkMultiplication signDepictionAndroid (robot)Virtual machineComputer-assisted translationMereologyRoboticsComputer animation
04:54
ComputerVirtual machineRoboticsMusical ensembleVirtual machineProjective planeMeeting/Interview
05:44
Machine visionObservational studyExpressionCondition numberProgrammer (hardware)RobotDrawing
06:30
ComputerCASE <Informatik>Radio-frequency identificationVirtual machineMultiplicationProgrammer (hardware)Moment (mathematics)Projective planeLecture/ConferenceMeeting/Interview
07:13
Projective planeMedical imaging
07:38
Artificial neural networkWebsiteVirtual machineRight angleRevision controlAlgorithmComputer animationPanel painting
08:16
Artificial neural networkCASE <Informatik>Digital photographyDifferent (Kate Ryan album)Level (video gaming)AlgorithmVirtual machineType theoryMedical imagingPhysical lawProgrammer (hardware)Decision theoryProjective planeNetwork topologyArtificial neural networkMeeting/Interview
09:12
ResultantMonster groupRight angle
09:40
Limit (category theory)MereologyGoodness of fitLimit (category theory)
10:02
Filter <Stochastik>Virtual machine
10:24
Computer-generated imageryData miningExt functorVirtual machineMathematical singularityAverageElement (mathematics)Projective planeVirtual machineAlgorithmSource code
10:59
Digital filterLevel (video gaming)Independence (probability theory)Machine learningParameter (computer programming)Decision theoryVirtual machineProgrammer (hardware)Perspective (visual)AlgorithmLevel (video gaming)AbstractionParameter (computer programming)Lecture/ConferenceMeeting/Interview
11:28
Physical lawParameter (computer programming)RoboticsPhysical lawDifferent (Kate Ryan album)Right angleAreaDependent and independent variablesVirtual machineBitChannel capacityMultiplication signDesign by contractComputer animation
12:16
Random numberMathematical singularityRobotRandomizationBitGroup actionProgrammer (hardware)Computational intelligenceRootRoboticsPurchasingVirtual machine
13:15
Digital photographyMeta elementPhysical lawAutomorphismInternetworkingStorage area networkAreaProfil (magazine)QuicksortPhysical lawSystem callHypermediaDigital photographyWeightMultiplication sign
14:09
Physical lawDecision theoryCopyright infringementCASE <Informatik>Digital photographyMeeting/Interview
14:47
Physical lawMusical ensembleModal logicOffice suiteImage registrationProduct (business)Order (biology)Link (knot theory)Price indexPhysical lawProcess (computing)Selectivity (electronic)WordAuthorizationOrder (biology)Mechanism designProgrammer (hardware)State of matterComputational intelligenceImage registrationRandomizationProduct (business)Different (Kate Ryan album)
16:32
Physical lawElement (mathematics)Personal digital assistantComa BerenicesStatement (computer science)Price indexAuthorizationPhysical lawMoment (mathematics)Virtual machineCASE <Informatik>WordDecision theoryComputer animation
17:35
Artificial intelligenceImage registrationProcess (computing)Virtual machinePerspective (visual)Decision theoryParameter (computer programming)CASE <Informatik>Programmer (hardware)Computer configurationMultiplication signRight angleWordAlgorithmComputer animation
19:05
Office suiteSound effectPhysical lawOrder (biology)Computational intelligenceMeeting/Interview
19:42
Artificial intelligenceImage registrationCASE <Informatik>Artificial neural networkRight angleAreaLecture/ConferenceComputer animation
20:41
Dependent and independent variablesRight angleRoboticsDecision theoryLine (geometry)Point (geometry)WindowType theoryArtificial neural networkMeeting/Interview
21:36
Control flowTheory of relativityLecture/Conference
22:20
FeedbackSummierbarkeitAreaMereologyTerm (mathematics)Artificial neural networkCopyright infringementLecture/ConferenceMeeting/Interview
22:40
Copyright infringementException handlingRight angleData miningDependent and independent variablesSubject indexingEmailAreaSystem callVirtual machineEndliche ModelltheorieInsertion lossMathematicsPhysical lawSheaf (mathematics)Lecture/ConferenceMeeting/Interview
24:15
Latent heatPoint (geometry)View (database)Lecture/Conference
24:49
Order of magnitudeArtificial neural networkLimit (category theory)AreaData structureBoundary value problemException handlingWhiteboardLogical constantLecture/ConferenceMeeting/Interview
26:33
TowerJSONXML
Transcript: English(auto-generated)
00:19
Thank you so much for staying. I know it's very close to the
00:24
closing of the conference, so I really appreciate it. I am from the University of Sussex, and I'm going to be talking to you about Androids and copyright. Now, I always start my presentations with an apology. I'm a lawyer. That's
00:47
what I'm talking about. I'm a lawyer, so I am going to be talking about art, and I'm going to be talking about artificial intelligence, and these are subjects that I'm not very familiar with. Now, the apology is, if there are any
01:05
people that are familiar with artificial intelligence, I apologise in advance in case I say something monumentally stupid. The reason why I think this is a very interesting subject is going to become evident later on, but also I'm
01:23
going to be talking a little bit about art, and also I'm quite interested in getting some feedback. If you are a German lawyer or if you are an artificial intelligence person that is interested in some of the issues of copyright that I'm going to raise, I do ask you for your help. Now, really,
01:46
what I want to start first with is with a few definitions and a few running through of what is artificial intelligence, and now this is obviously an entity that has a
02:03
capacity to perceive, understand, predict and manipulate the world far larger and more complicated than itself, and this is one of the traditional definitions of what is artificial intelligence. By the way, there is a lot of pictures of Star Wars, happy Star Wars Day, may the fourth
02:22
be with you. Now, two concepts that I've been quite interested in that I think are generating a lot of potential legal interest is the subject of neural networks which we've heard quite a lot, and yesterday I heard an amazing presentation on this very subject. Now, neural
02:43
networks, you will know already, are networks that try to mimic human thoughts, and particularly with basic mathematical models of a neuron. But also I'm quite interested in machine learning, and machine learning is usually
03:02
defined as a computer that is going to have the ability to learn without assistance by the programmer. Now, what has been happening that I think is raising legal questions is in the matter of artificial intelligence art. Now, Eva, for
03:23
those of you who have seen Ex Machina, it's a fantastic film to begin with, but also it contains a very interesting scene in which Eva, the artificial intelligence, the robot pictured here, starts drawing one of the other
03:44
protagonists, and this is presented usually as a trope in science fiction to tell us that the robot, the artificial intelligence, is striving to be a human, more human, is showing intelligence. This is, as I was saying, a
04:04
trope. It's handled in quite a lot of depictions of artificial intelligence in film and television. If you're familiar with Star Trek, you will notice that this picture, this cat, is actually painted by Data in Star Trek, the
04:22
first generation, and Data is very, very strongly a robot, an android, that wants to become more human, wants to be more human, and he does this through art. And we are presented a lot of times with this trope, the idea that art
04:43
made by robots is actually going to be part of their humanity, is going to create a more human machine. And we see this time and time and time again in science fiction. So the robots are going to create photography, are going
05:00
to create music, are going to create poetry, are going to create literary works that express their humanity. Now, this is actually something that has been happening quite a lot. This is a painting in the 60s of art, machine art by
05:22
Desmond Paul Henry, and it actually shows that this is something that has been happening for many years. Back in the 70s and 80s, there were other projects, things like Aaron, E. David, the painting for Ava I'm going to show, and more recently, these things have been getting more
05:41
interesting with Deep Dream and the next Rembrandt that I'm going to show. So this is a painting by E. David, and it's a criticism, an article that is actually telling us that the artist itself, the artist is the expression of a portrait, the expression, it's an expression of the
06:01
personality of the artist. And I find that quite compelling, that I'm going to tell you the legal definition of what is creativity, but that art critics are actually looking at this as the expression of the programmer. Now, it documents the human condition, as Savannah
06:23
Cox says, but which human condition is the human condition of the programmer, of the artist? Is it of us who are looking at these beautiful paintings that are created by machines? Some of these have grand aspirations. This is the painting full, telling us that it wants to be
06:41
taken seriously as a painter, and this is a series of paintings that is done on Amelie. We are also going to have the idea at the moment that artists are becoming something more. So the early artists were very dependent on
07:03
what the programmer told them to do. So this is a painting by a long project called Aaron, that I find quite interesting. And so, Eva, sorry, Eva, this is a good
07:20
example. If you have ever seen X Machina, it's trying to represent by a filter the art that is drawn in the picture by Eva. This is actually a project that you can do. Apologies for the next image. This is actually me
07:40
taking a picture. It's a website, and you can go to the website, and you can take a picture, or you can upload a picture, and it will convert it into machine art. And this is going this is just a filter, right? There is not artificial intelligence happening. But what's interesting is that the newer versions, things like
08:02
Deep Dream, and I've seen quite a lot of talk already about things like this and Deep Dream, is that through machine learning, there is going to be an algorithm that is going to be selected by the machine itself. So you take a picture, and the machine is going to apply
08:22
certain types of algorithms that is going to create an image that was not specifically designed by the programmer. So really what is going on is that instead of exactly prescribing and this is a quote directly from the Deep Dream project, instead of exactly prescribing
08:44
which feature we want the network to simply to amplify, we can also let the network make the decision. And this is quite interesting. That is beyond all of the electronic artists that we were seeing before. These artists are actually making the decisions themselves
09:03
through decision trees and things that I don't fully understand. But they pick layers, and they are making some decisions, and this is quite important for the law. Now, you can do it yourself. This is a picture that I took in on the one side. It's a picture I took from Lake
09:23
Cuomo, and then I put it through one of the many places where you can do alteration with Deep Dream, and the result of it looks like a monster destroying Italy or something like that. And artists and art commentators
09:44
are saying, were asking the question, is this art? And actually it is. This is Ben David that says, of course it's art. There is no limit to what you can classify as art. The question is only ever whether it's good art, and people seem to be very amused by it. So at least it
10:01
produces amusement. So we have more than just the creation of filters. The machines, you can tell them, you have a photograph, and apply the painting of Rembrandt or the cream, and it can create something that resembles the style of the painter. So is this
10:21
creation or is this imitation? And that is quite interesting. And this is the latest that you probably have seen it if you're interested in the subject. The next Rembrandt is a Microsoft research project that actually went through all of the paintings of Rembrandt to find
10:40
through machine learning algorithms, elements, common elements in all of the paintings, and created a new painting that did not exist, that is completely, according to some critics, in the style of Rembrandt. Now, this, what is happening is something that we
11:01
don't really know from a legal perspective. I'm going to tell you. But it depends on what is really, really happening behind. So my understanding of the decisions is that the programmers create some algorithms, and the machine is actually making quite a lot of the decisions that are going to be showing us new
11:21
art. So preselected levels of abstraction that are going to be dependent on certain parameters. So what does the law tell us about what is going on? I had a different picture here, but I was so taken by the robot wall outside that I decided to put it here.
11:42
Now, this is not a new area of law. We have been, since the Roman times, dealing with things like law of slavery. It's rights and responsibilities of those who
12:01
do not have legal capacity, and from slaves to minors to people who are mentally impaired. This is not an area that should give us a lot of difficulty. And we have areas like negligence, contract, tort, that have ways of dealing with this. Machines are doing a little bit
12:20
more. This was presented last year here at the Republica, which is a robot that goes, a computer, I guess, yeah, a robot. It's called a random dark net shopper. It goes to the dark net, and with a certain
12:42
amount of bitcoins, it's going to purchase random things from the dark net, and it purchases drugs, and you can find them here. Actually, the Swiss police confiscated the computer, and then returned it saying, well, you know, the artistic value of this is more
13:03
important than the possible damage that was created. So interesting things are happening because the programmers didn't tell it go on by drugs, but it actually just went and randomly bought drugs from the dark net. Then I'm sort of specializing in this area
13:21
of law recently, which is what you could call non-human copyright, and there's also an increase in this area. I don't know if you're familiar with this guy. He is Naruto. He is a monkey in Indonesia. But in 2011, a British photographer put a camera, and
13:42
the monkey took a selfie. Now, it would be a fun story in itself. For a time, I think back in 2012, everyone was changing their social media profiles to the monkey, and it was very, very cute. I mean, he's a very handsome monkey, you have to admit. Now, what is interesting is that there was a copyright
14:02
lawsuit in San Francisco based on the monkey. Now, PETA, the People for the Ethical Treatment of Animals, sued the photographer for copyright infringement, arguing on behalf of the monkey, arguing that the monkey took the pictures, so the monkey should have the copyright.
14:22
So this, of course, the judge came to the decision that the monkey did not have a standing in court, so ruled against the claimants, but this is a subject that is coming up again and again and again, and I'm interested, and I think that there may be a case for something
14:42
called non-human copyright, and if you're interested in the monkey selfie, I know you cannot see it because I'm covering it. There is a link, I wrote a paper telling you all about the copyright implications of the monkey selfie. Now, I'm going to disappoint you, because
15:03
UK law is actually quite boring in this respect. What the law says, and as far as I know, this is the only country, the UK, where something like this
15:21
is legislated, and I would be delighted to hear if any other jurisdiction has something different. Now, UK law has the concept that the programmer of any computer generated work of art is going to be the owner of the
15:41
copyright. It's very simple, and it's, you can call it the automated, the computer-generated copyright. Now, the US Copyright Office, which handles registration in the United States, has a different approach, and there they say, this is just a guideline, but generally the
16:03
courts in the United States have been following in this. They're saying that in order to be entitled to copyright registration, a work must be the product of human authorship, so works produced by mechanical processes or a random selection without any contribution by a human
16:21
author are not registrable, et cetera, et cetera. So if it's mechanical, if it's not produced by a human, it's not going to be registrable, and therefore you cannot enforce your copyright. European law is slightly different, and this is actually why, what prompted me to write this paper and start doing the research, because
16:41
European copyright law has been moving very strongly to the protection of the personality of the author, so actually the wording of the case law and the wording of the law says that the work must reflect the author's personality, and this is very strongly that is coming across in several cases, so that's Painter,
17:04
Infopack, and others. So even though it's unclear at the moment, I think that we have quite strong indication that works that do not reflect the personality of the author, therefore they're created by a machine, and the
17:25
creative decisions are made by the machine are not going to be protected by copyright, and this is, I know it's a very strong statement, but I believe it. So this is where I think we are finding ourselves in very
17:41
interesting territory from legal perspective. If we accept that the very important, very basic creative decisions in the creative process that are being made by machines, by the artificial intelligence, by the machine learning
18:01
algorithms, very important decisions that we don't, we cannot assign to the programmer, if those decisions have been made by the machine, there is a very strong chance, and I'll be very willing to talk to you about this, but I think that there is a very strong argument
18:24
to be made that in Europe, and almost certainly in the United States, those works would not be protected by copyright, and that maybe doesn't have any implication at the moment, but the value of some of these works may increase with time. The next Rembrandt, I have no
18:42
idea how much of an investment came into that, but I would actually say that the programmers made quite a lot of decisions in that circumstance, so I would say that probably that particular case would be protected by copyright, so we have several options. Either those works do not have copyright, and we forget about it,
19:03
you know, it's a nice experiment, it's not going to be protected by copyright, they're not going to be allowed to be registered, possibly in the US Copyright Office, again, this would affect works in the United States, and it could be possible. Something that probably
19:23
I would like is to actually, for us to apply the UK computer-generated clause to a more wider audience, so it could actually become European law, and I don't know, some countries could change their legislation in order
19:40
to implement this. Or we could take a wider approach and start recognising non-human copyright. Now, there is an interesting case, just to tell you just how much non-human copyright we're actually talking about, there
20:00
is a very interesting case in the United States called Urantia Books, versus Mashera, and this case is about alien-dictated copyright, not only alien-dictated copyright, but dead alien channeling into a person, so
20:22
this is really the definition of non-human copyright, and that work actually was granted copyright, because there was a human that was transcribing everything, but it's just an interesting area, I just find the monkeys and artificial intelligence and alien copyright is really my subject, you know. So there could be maybe the
20:45
creation of a new type of copyright dealing with this, or maybe we should start thinking ahead in the future, and I'm going completely into the science fiction gazing here, start thinking of non-human rights and
21:04
responsibilities along the lines of what happens with a robot, artificial intelligence buys drugs, what happens when it kills someone, which is going to happen, and what happens when it creates a work of art. I don't think that robots are actually going to be asking for copyright
21:21
tomorrow, they're not going to rise up, and if they rise up and kill us at some point, I don't think that copyright is actually going to be very high in their agenda, in the decision to do that, I would blame Windows 10. But anyway, so with those thoughts, thank you
21:40
very much for your attention. Yeah, thank you very much. Thank you very much for this talk. We'll have a short break, 10 minutes. We have 10 minutes left, so does anybody have questions in the room?
22:03
God, I would be surprised. Now there are no questions or comments, maybe? Something Star Wars related? Yeah. If I say something absolutely stupid and artificial intelligence, I do want to hear it, honestly. I welcome
22:22
the feedback. It's an area that I really need to learn more about. Yeah. Hi, thank you for the talk. I'm just curious about, so you talked about a lot about the creation, copyright in terms of art creation by artificial intelligence, but have you also looked at copyright infringement by the
22:42
AI? Is that a topic that you're looking at? That's an excellent question. I think it's already happening in a certain way. A lot of, there is a lot of automatic indexing that, I know, we actually have to create exceptions to copyright because of some of
23:04
these things that happen mechanically. So actually, if there is a machine that is doing something that is large-scale copyright infringement, there may be an exception on that. There is an area that now the UK has adopted, which is with data mining, because
23:23
a lot of data mining is actually being done, it's copyright infringement. You need to create a copy, usually lift a copy of the work and create a copy of that work. So we actually have to change UK law to create a data mining exception that allows for
23:41
research purposes for a machine to do this work, which is actually copyright infringement. I think it's interesting. I haven't looked at it as such, but now you've given me an idea. I'm going to have to include a section, and please send me an email or
24:01
something, and I'll cite you in the article, and it's finally published. But no, it's fascinating. Yeah, I think it should also include, if we're talking about rights and responsibilities, it should definitely include a section on infringement. Thank you very much. I thought it was a very
24:24
interesting, crazy idea of a presentation, so I really appreciate that. On the other hand, I'm wondering if we should continue talking about copyright as a matter of protecting creative works from a capitalist point of view, and rather than
24:43
talk, if we're talking about the future, about more open and less restrictive way of protecting works, I'm thinking about the TPP, the TTIP, all these international treaties that are being pushed and that do limit creation and
25:01
innovation in a way. So I was wondering what are your thoughts on this, if we could shift from copyright to creative commons and artificial intelligence. Thanks very much. That absolutely falls in my area of interest. I actually do a lot of work for creative commons. I was creative commons Scotland, creative
25:21
commons Costa Rica, now currently legal lead for creative commons UK, so you probably guess where I land in this debate. I'm all for reform. I'm
25:40
interested in reform. I'm interested in reform. There's one thing that we can do without copyright, at least not on the current political structures. We'll have to live with copyright at least for quite a while. But yeah, I completely agree. We should be moving into more of these more permissive areas, and that's actually one of the areas that I'm interested in and looking at, that if we can reform
26:01
copyright to think outside of the boundaries of generally that we have been looking at them in the past, like just looking at what I call non-human copyright, we would actually I think be looking more at creating more exceptions
26:20
and more public related exceptions that I think we can use in the future. And I'm for copyright reform and copyright reform all across the board. Yeah, thanks. Thanks for the question actually. Is there a last question in the audience? No. So thank you very much. Okay, thanks.