Download as PDF, PPTX
![CAN AI BE AN INVENTOR?
Stephen L Thaler v. The Comptroller-General of Patents,
Designs and Trade Marks
[2020] EWHC 2412 (Pat)
• UK Patent application GB1816909.4 – Food container
• UK Patent application GB1818161.0 – Search and
Rescue Beacon
• Inventor: [DABUS]
• Filed by: Dr. Stephen Thaler](https://image.slidesharecdn.com/agsaiipsnufdpfinal-201018170450/85/Artificial-Intelligence-and-Intellectual-Property-33-320.jpg)
![CAN AI BE AN INVENTOR?
Stephen L Thaler v. The Comptroller-General of Patents,
Designs and Trade Marks
[2020] EWHC 2412 (Pat)
• UK Patent application GB1816909.4 – Food container
• UK Patent application GB1818161.0 – Search and
Rescue Beacon
• Inventor: [DABUS]
• Filed by: Dr. Stephen Thaler](https://image.slidesharecdn.com/agsaiipsnufdpfinal-201018170450/75/Artificial-Intelligence-and-Intellectual-Property-33-2048.jpg)
Uploaded byArul Scaria
Artificial Intelligence and Intellectual Property
The document discusses the intersection of artificial intelligence (AI) and intellectual property (IP) rights, highlighting how AI can both aid in IP management and challenge traditional IP frameworks. It raises critical questions about the ownership of AI-generated works, the legal status of AI as an inventor, and the potential need for reform in IP law to accommodate the unique challenges presented by AI technologies. Ultimately, it calls for a reevaluation of IP protection, emphasizing the importance of human involvement in AI outputs and proposing stronger disclosure requirements.
Law◦
More Related Content
Similar to Artificial Intelligence and Intellectual Property
![AI and the Future of Work [TUG-CO, 11/15/23]](https://cdn.slidesharecdn.com/ss_thumbnails/aiandthefutureofworktug-conov152023-231116175445-e711e59c-thumbnail.jpg?width=600ounds&width=560&fit=bounds)
Artificial Intelligence and Intellectual Property
- 1. ARTIFICIAL INTELLIGENCE AND INTELLECTUALPROPERTY Dr Arul George Scaria Faculty Development Programme, Shiv Nadar University, October 14, 2020
- 2.
- 3. AI IS EVERYWHERE! Imagesource: https://www.gettingsmart.com/2018/12/the-future-is-here-artificial-intelligence-what-it-means-for-our-kids-2/
- 4. INTERFACE OF IPWITH AI • AI as a tool for IP management • IP as a tool for protecting AI/ AI generated works • IP as an obstacle to innovation and transparency in AI
- 5.
- 6. AI as atool for IP management
- 7. AI & IPMANAGEMENT • Useful tool in administration and processing of IP applications • Document verifications • Prior art search • Image search • Translations of documents • … • Challenges?
- 8. IP as atool for protecting AI/ AI generated works
- 9. IP AS ATOOL FOR PROVIDING EXCLUSIVE RIGHTS IN- AI AI generated works ?
- 10.
- 11. CONTEXT • Bloomberg -‘Cyborg’ - company earnings reports • The Associated Press - Minor league baseball • The Washington Post - High school football • The Los Angeles Times - Earthquakes • Forbes - Bertie - rough drafts and story templates Source: https://www.nytimes.com/2019/02/05/business/media/artificial- intelligence-journalism-robots.html
- 12.
- 13. CONTEXT Google, Stanford Universityand University of Massachusetts - Recurrent Neural Network Language Model (RNNLM) Source: https://www.theguardian.com/technology/2016/may/17/googles-ai-write-poetry-stark-dramatic- vogons
- 14. CONTEXT Deepbach https://www.youtube.com/watch?v=QiBM7-5hA6o Flow Machines project(ERC funded project), led by Dr François Pachet
- 15.
- 16. CONTEXT Bayou • Uses codesposted on GitHub • Neural sketch learning - associates an “intent” behind each code • Creates apps by matching the user requests to intent Image source: https://pixabay.com/photos/code-coding-web-development-944499/
- 17. CONTEXT US Patent Number:US010770180 Date of grant: September 8, 2020 Title: Processing clinical notes using recurrent neural networks Abstract: Methods, systems, and apparatus, including computer programs encoded on computer storage media, for predicting future patient health using neural networks. One of the methods includes receiving electronic health record data for a patient; generating a respective observation embedding for each of the observations, comprising, for each clinical note: processing the sequence of tokens in the clinical note using a clinical note embedding LSTM to generate a respective token embedding for each of the tokens; and generating the observation embedding for the clinical note from the token embeddings; generating an embedded representation, comprising, for each time window: combining the observation embeddings of observations occurring during the time window to generate a patient record embedding; and processing the embedded representation of the electronic health record data using a prediction recurrent neural network to generate a neural network output that characterizes a future health status of the patient. http://patft.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=%2Fnetahtml%2FPTO%2Fsearch- adv.htm&r=1&f=G&l=50&d=PTXT&S1=10770180&OS=10770180&RS=10770180
- 18. CONTEXT Kit-Kay Mak andMallikarjuna Rao Pichika, ‘Artificial Intelligence in Drug Development: Present Status and Future Prospects’ Drug Discovery Today (Vol. 24(3), 2019), 773-780.
- 19.
- 20. THREE IMPORTANT IPRIGHTS IN THE CONTEXT OF AI Copyright Patents Trade Secrets
- 21. Are there anyjustifications for grant of IP rights over AI or outputs generated by AI? What can we learn from theories of IP?
- 22. WELFARE THEORY ‘Incentive Theory’ Lawand economics approach Roots – utilitarianism
- 23. WELFARE THEORY Information goods- ‘Public goods’ characteristic Non-rivalrous character in consumption Relative non-excludability Potential consequence?
- 24. WELFARE THEORY Possible solutions: Governmentprovides the good Government subsidises select private innovators Government issues prizes to successful private producers Legal reinforcement of self-help strategies Government protects producers against competition
- 25. FAIRNESS THEORY Classic form:Locke Everyone has a natural right to the fruits of her/ his (intellectual) labour Resources that are unowned or “held in common” v. Facts/ concepts Image source: http://newint.org/archive/images/issue/376/pics_ken1.gif
- 26. FAIRNESS THEORY Modified form:Equity theory Applications of the theory?
- 27. PERSONALITY THEORY Kant andHegel Personhood theory of ‘real property’ - Private property rights required to satisfy some of the basic human needs - Policy makers should try to create and allocate entitlements to resources so that people can fulfil those needs Image source: http://www.freemarketfoundation.com/DynamicData/Articles/Images/8154/prop.jpg
- 28. PERSONALITY THEORY Extension toIP - Intellectual works are manifestations or extensions of the personalities of their creators - IP rights can shield from misappropriations and modifications
- 29. CULTURE THEORY ‘Social PlanningTheory’ Roots in different disciplines - Writings of Marx, Sen, Kahneman, Benkler, Fisher, etc. Property rights - IPR - can and should be shaped to help in fostering the achievement of a just and attractive culture
- 30. CULTURE THEORY Existence of‘human nature’ - Bloom more under certain conditions - Adapt social and political institutions to facilitate that blooming Conditions necessary for the full realization of personhood Life, Health, Engagement, Privacy, etc. Access to these conditions Currently - limited and systematically unequal Should try to expand and to some degree equalize access Implications for IP policy? Image source: https://www.economist.com/united-states/2011/09/17/pinched
- 31. • Do thesetheories support grant of IP rights over AI? • Do any of these theories support grant of IP rights to AI over the outputs generated by AI? • Should the approaches be different for Weak AI and Strong AI?
- 32. SOME OF THEINEVITABLE QUESTIONS IN THE CONTEXT OF PATENTS Can AI be considered as an ‘inventor’? Who will be the owner of a patent, wherein AI is the inventor? If only humans can be considered as ‘inventors’, who should be the inventor? Should we revisit the inventive step/ non-obviousness tests? Who will be the person skilled in the art? How would we meet the disclosure requirements, particularly when the code change over time in an autonomous manner? (“AI Blackbox”) How would we address the challenges in claim drafting? Should there be mandatory disclosure of training datasets as part of patent application? How do we address the question of patent infringements by AI?
- 33. CAN AI BEAN INVENTOR? Stephen L Thaler v. The Comptroller-General of Patents, Designs and Trade Marks [2020] EWHC 2412 (Pat) • UK Patent application GB1816909.4 – Food container • UK Patent application GB1818161.0 – Search and Rescue Beacon • Inventor: [DABUS] • Filed by: Dr. Stephen Thaler
- 34. STEPHEN L THALERV. THE COMPTROLLER- GENERAL OF PATENTS, DESIGNS AND TRADE MARKS • Key argument of Dr Thaler: Inventorship not confined to natural persons • UK IPO: DABUS was a machine - cannot be considered as an inventor under the Patents Act 1977 - Only a natural person can be an inventor • HC (Jus. Marcus Smith): A patent could only be granted to a person – There is an established principle that only a person can hold property rights
- 35. ANALYSING THE CHALLENGESIN THE © CONTEXT • Can AI be the author of outputs generated by AI in an autonomous manner (deep learning)? • Should we revisit the ‘originality’ standards in copyright law?
- 36. CAN AI BEAN AUTHOR? Insights from Naruto v. Slater 888 F.3d 418 (9th Cir. 2018)?
- 37. ARGUMENTS IN FAVOUROF GRANTING PROTECTION • AI can generate value and someone should be able to capture that value • Free competition between human created and AI created outputs requires that AI outputs be given the same kind of protection given to humans • The programmer/ owner/ user of an AI machine can be considered as the ‘proxy author’ for copyright protection purposes – Applicability of WFH doctrine?
- 38. ARGUMENTS AGAINST GRANTING PROTECTION •Would providing IP protection to AI outputs incentivise creativity or innovation? • Would there be over-rewarding of programmers, particularly when the outputs autonomously generated by AI were unforeseeable/ unpredictable? • Copyright and patent laws are historically centered around human beings • Limitations of WFH doctrine in the AI context • If granted rights, will the owners/ users of AI be willing to take responsibilities for copyright infringement, patent infringement, libel, etc.?
- 39. IP as anobstacle for innovation and transparency in AI
- 40.
- 41. IP AS ANOBSTACLE FOR INNOVATION AND TRANSPARENCY • The cost of lack of transparency in AI • IP rights as a hindrance to transparency • Algorithmic collusion and competition • Would protection for AI promote or defeat dynamic efficiency?
- 42. CONCLUSION • IP protectioninvolves social costs - If IP rights are not clearly providing a benefit, there is hardly any justification for the social costs imposed by IP protection • Make IP protection contingent on human input linkage - If human intervention in the creation of the AI outputs cannot be established, no protection should be given under IP laws • If patent protection is given for AI outputs, strong disclosure requirements are required – public deposit of algorithm and training datasets • Need for broader copyright exceptions – TDM exceptions
- 43.