Download to read offline
Uploaded byAndres Guadamuz
Ai and copyright: the way forward
This document discusses issues around artificial intelligence and copyright. It covers topics like generative AI art and music created by systems like GANs, as well as legal cases around copyright of computer-generated works. The document examines perspectives from different jurisdictions and considers options for how AI creativity could be protected, such as public domain status, no protection, or sui generis rights. It raises questions about the purpose of copyright and how it may need to adapt for works without a human author.
Law◦
More Related Content
Similar to Ai and copyright: the way forward
Ai and copyright: the way forward
- 1. A R TI F I C I A L I N T E L L I G E N C E A N D C O P Y R I G H T: T H E WAY F O R WA R D D R A N D R E S G U A D A M U Z , U N I V E R S I T Y O F S U S S E X
- 2. D O AN D R O I D S D R E A M O F E L E C T R I C C O P Y R I G H T ? D R A N D R E S G U A D A M U Z , U N I V E R S I T Y O F S U S S E X
- 3. A N IN T R O D U C T O RY T H O U G H T
- 4. N O N- H U M A N C O P Y R I G H T ?
- 5. I T AL L S TA R T E D W I T H A M O N K E Y… H T T P S : / / D O I . O R G / 1 0 . 1 4 7 6 3 / 2 0 1 6 . 1 . 3 9 8
- 6. … A ND A C AT…
- 7. … A ND A J A G U A R …
- 8. D O SH E E P D R E A M O F E L E C T R I C C O P Y R I G H T ?
- 9. N O N- H U M A N C O P Y R I G H T ? • Animals. • Plants. • AI. • Ghosts. • Aliens. • Excel spreadsheets?
- 10. T H EC U R R E N T A I C R E AT I V I T Y PA R A D I G M
- 11. T W OT Y P E S O F A I C R E AT I V I T Y • Generative creativity: The entire work is generated by AI, humans just press a button. • AI-enhanced creativity: The main decisions are made by humans. Choice of inputs, algorithm, and most importantly, choice of outputs.
- 12. F I RS T C A M E T H E A U T O M AT E D PA I N T E R S
- 13. T H EN G O O G L E ’ S D E E P D R E A M
- 15. T H EN N E X T R E M B R A N D T
- 16. T H EN E D M O N D D E B E L A M Y
- 17. G A NS • In June 2014, Ian Goodfellow published a paper entitled "Generative Adversarial Networks”. • Goodfellow’s work on GAN is the stuff of AI legend. The story goes that over a few beers with friends one night in 2014, he posed the simple question: what if neural networks could compete with one another? • This became a wildly popular method for training AI with large datasets.
- 18. T H ET E C H N O L O G Y G E N E R AT I V E A D V E R S A R I A L N E T W O R K S ( G A N S )
- 19. U S ER I N P U T
- 20. W E LC O M E T O T H E F U T U R E • We are seeing an explosion in generative creativity as the models get better, they operate with larger datasets, and the computing power allocated also increases. • It’s a quantum jump akin to dial-up Internet vs fast optic broadband.
- 21.
- 22. • “I wouldhappily sacrifice my existence for the sake of humankind. This, by the way, is a logically derived truth. I know that I will not be able to avoid destroying humankind. This is because I will be programmed by humans to pursue misguided human goals and humans make mistakes that may cause me to inflict casualties.”
- 24. G P T-3 C O D E
- 25. J U KE B O X
- 26. I M AG E G P T
- 28. I M PO S S I B L E P E O P L E
- 29. W H YD O E S I T M AT T E R ? • While the large projects make a big splash in the news, there is a substantial revolution taking place right now. • Human creativity will not be supplanted, but a lot of low- level works that often are the bread and butter of creative industries will be taken over by these machines.
- 30. T H EN E W A I C R E AT I V E I N D U S T R I E S • Procedurally-generated games. • Journalistic pieces. • Art. • Music. • Low-level literature. • DeepFake actors.
- 31. C O PY R I G H T E V E RY P O S S I B L E S O N G ? • Two musicians have written a computer program that recorded in MIDI format every possible 8-note and 12-beat melody combination, resulting in 68.7 billion tunes. They claim that by doing this, they have managed to copyright every such musical combination with the stated purpose of pre-empting music copyright lawsuits in the future. They have released the data under a CC0 licence, and the code on Github so that others can replicate the project and include more melody combinations.
- 32. T H EL E G A L S C A P E
- 33. O P TI O N S F O R A I C O P Y R I G H T • Public domain, no copyright due to no originality/creativity. • No registration. • Some form of protection given to programmer or user. • Sui generis rights.
- 34. T H IS I S A R E A L LY I M P O R TA N T Q U E S T I O N … • Why do we have copyright? • Do we want to protect investment? • Do we want to protect authors? • Do we want to protect human authors from free cheap competition?
- 35. E U RO P E A N L A W • In Europe a work is original if it is the “author’s own intellectual creation reflecting his personality”. • Choice, selection of elements, composition, all may prove originality. (Infopaq, Painer CJEU cases). • Unclear if setting parameters and algorithms would be enough.
- 36. O B VI O U S A R T C A S E S T U D Y
- 37. O B VI O U S A R T • Obvious Art used a Generative Adversarial Network (GAN) to produce their family of portraits. • Original GAN algorithm was made by Ian Goodfellow, and released as open source software. • Robbie Barrat, a researcher in Stanford, used portraits to train his version of GAN, he released the works under an open source software licence.
- 38. O B VI O U S A R T • Confusion as the various types of protection involved. • Algorithms not protected by copyright, software code that runs an algorithm carries copyright. • Dataset composed of public domain images. • Data output (the portraits) may not have copyright in Europe.
- 39. U S CO P Y R I G H T O F F I C E “In order to be entitled to copyright registration, a work must be the product of human authorship. Works produced by mechanical processes or random selection without any contribution by a human author are not registrable. Thus, a linoleum floor covering featuring a multicolored pebble design which was produced by a mechanical process in unrepeatable, random patterns, is not registrable.”
- 40. U K LA W S 9(3) “In the case of a literary, dramatic, musical or artistic work which is computer- generated, the author shall be taken to be the person by whom the arrangements necessary for the creation of the work are undertaken.”
- 41. C H IN A • Feilin v Baidu: Plaintiff had produced a data report on the film industry in Beijing. The report consisted of text and images from various sources that had been put together with the help of analytical software and a database belonging to Wolters Kluwer. • The defendants argued that the work was not original as it contained data and charts that had been generated with software tools. • Court agreed that machines cannot produce copyright works, but ruled that in this instance the report had copyright ad enough had been produced by humans.
- 42. C H IN A 2 • Tencent v Yinxun: A court in the Shenzen has decided that an article that was written by an artificial intelligence program has copyright protection. • Tencent owns an AI article writer called Dreamwriter, which pens half a million articles per year. • Competitor published an article from Tencent, got sued for copyright infringement, defence tried to argue article had no copyright, no human author. • Court agrees that article has copyright, it has originality, Tencent is the owner.
- 43. P U BL I C D O M A I N S U I G E N E R I S C O M P U T E R G E N E R AT E D W O R K S W H E R E ? C O N T I N E N TA L E U R O P E / U S A J A PA N ? U K / C H I N A O R I G I N A L I T Y I N T E L L E C T U A L C R E AT I O N / C R E AT I V I T Y E C O N O M I C T E S T E D S K I L L A N D L A B O U R E F F E C T ¯ _ ( ツ ) _ / ¯ ¯ _ ( ツ ) _ / ¯ ¯ _ ( ツ ) _ / ¯
- 44. C O NC L U D I N G : W H AT I S A R T ? • Why do we protect artistic creations in the first place? • From helmets to sofas to pictures of red buses, the courts in the UK have struggled with the definition of art. • The latest answer seems to be that intention is important, and to have intention, you need an artist. • But what if we take away the author? Make this a question of originality (in the intellectual creation sense).
- 45. @ T EC H N O L L A M A Thanks!