인공지능을 발명자로 볼 수 있을까? -DABUS 판결에 부쳐 = Can Artificial Intelligence be Recognized as an Inventor? -In Re DABUS Cases

다국어 초록 (Multilingual Abstract)

Decisions on patent applications referring to DABUS, artificial neural network (ANN), as the inventor are being released. The UK's High Court of Justice and the Court of Appeal, the US District Court for the Eastern District of Virginia, and the Federal Patent Court of Germany (Bundespatentgericht) ruled that the inventor of an invention should be a natural person in terms of respective patent law in force. Meanwhile, the Federal Court of Australia held that artificial intelligence (AI) could also be identified as an inventor under the Patents Act 1990 of Australia. All of these judgments relied on normative, deductive interpretation.
However, the patent law is not a fixed norm, as it reflects a strong policy perspective. Therefore, whether AI can be recognized as an inventor cannot be addressed a priori. When it is not studied what specific roles the AI, designated as the inventor of the relevant invention carried out, judgments might be made based on impressions, superficial understandings, and fictional premises, far from the factual ground.
ANNs are optimization models that derive desired output from input data through matrix operations, i.e., long sequences of summations and multiplications. Notwithstanding the complexity of “automated” computation, such mathematical calculation should not be equated with an “autonomous” invention of AI. At least under state-of-the-art technology, a machine does nothing but computes an optimal solution based on given data, following the instructions programmed by humans. From the technological point of view, AI can hardly be approved as an inventor yet. There is not much need to do so, even from the normative and policy perspective up to now.

국문 초록 (Abstract)

인공신경망 DABUS를 발명자로 기재한 특허출원에 관한 판결이 하나둘 선고되고 있다. 영국의 제1, 2심과 미국, 독일의 제1심판결은 특허법 해석상 발명자는 자연인을 지정해야 한다고 보았다. 반면 오스트레일리아 연방법원은 인공지능도 특허법상 발명자가 될 수 있다고 판결했다. 이들 판결은 모두 자국 현행법의 연역적 해석에 크게 기대고 있다.
그러나 특허법은 정책적 성격이 강한 법으로, 고정된 규범이 아니다. 따라서 ‘인공지능을 발명자로 볼 수 있는가’하는 문제 또한 선험적으로 답할 수 있는 문제가 아니다. 특허출원의 발명자로 기재된 인공지능이 해당 발명에서 구체적으로 어떠한 역할을 했는지가 밝혀지지 아니한 상태에서는, 인공지능에 관한 상상 속 이미지와 불완전한 이해, 허구적 전제를 바탕으로 실제와는 동떨어진 판단을 내릴 수 있다.
인공신경망은 입력된 데이터로부터, 덧셈과 곱셈에 기초한 행렬 연산을 통해, 원하는 출력에 이르기 위한 최적의 파라미터를 찾아내는 수학적 모형이다. ‘자동화된’ 수학적 계산이 아무리 복잡하더라도, 이는 인공지능이 ‘자율적으로’ 발명하였다는 것과 동일시될 수 없다. 적어도 현재까지 나온 기술 수준에서, 기계는 인간이 짠 최적화 알고리듬에 따라 주어진 데이터를 바탕으로 해를 찾을 뿐이다. 따라서 기술적 측면에서 볼 때 인공지능은 아직 발명자로 보기 어렵다. 규범적, 정책적 측면에서도 현재로서는 인공지능을 발명자로 인정할 필요가 크지 않다.

참고문헌 (Reference)

• 1 권보원, "회사를 처벌하는 양벌규정의 실효성에 대한 비판적 검토 - 회사불법 억지를 위한 최적 법제 모색 시론 -" 한국법경제학회 15 (15): 197-258, 2018
• 2 김선희, "호주의 입헌주의와 연방대법원의 역할" 헌법재판연구원 2020
• 3 최승재 ; 이진수, "헌법 제22조와 발명자권, 특허법개정에 대한 연구" 한국지식재산연구원 16 (16): 31-70, 2021
• 4 Bishop, Christopher M., "패턴인식과 머신러닝" 제이펍 2018
• 5 특허청, "특허청, 인공지능을 발명자로 기재한 특허출원에 보정요구서 통지" 2021
• 6 정상조, "특허법 주해Ⅰ" 박영사 2010
• 7 조영선, "특허법 3.0" 박영사 2021
• 8 윤선희, "특허법" 법문사 2019
• 9 특허청, "특허·실용신안 심사기준(2021. 12. 30. 개정 특허청 예규 제124호)"
• 10 Kurzweil, Ray, "특이점이 온다: 기술이 인간을 초월하는 순간" 김영사 2007
• 11 Chollet, François, "케라스 창시자에게 배우는 딥러닝" 길벗 2018
• 12 정상조, "지식재산권법" 홍문사 2020
• 13 윤영진, "지식재산권 보호 정당성 고찰을 통한 인공지능 창작물의 새로운 보호 방안 연구" 16 (16): 2020
• 14 Hutter, Frank, "자동머신러닝" 에이콘 2021
• 15 윤길준, "인공지능이 한 발명에 대한 특허" 2018
• 16 김광남, "인공지능시대를 맞이하는 특허법상 도전과 혁신 - 인공지능이 한 발명의 법적취급을 중심으로 -" 한국법학원 182 (182): 187-224, 2021
• 17 조영선, "인공지능과 특허의 법률문제" 법학연구원 (90) : 197-231, 2018
• 18 전정현 ; 김병필, "인공지능과 법률 서비스:현황과 과제" 한국법학원 170 (170): 218-258, 2019
• 19 한국인공지능법학회, "인공지능과 법" 박영사 2019
• 20 계승균, "인공지능(AI) 분야 산업재산권 이슈 발굴 및 연구" 특허청 산업재산정책과 2016
• 21 김시열, "인공지능 등 비자연인의 특허권 주체 인정을 위한 인격 부여 가능성에 관한 연구" 법학연구소 39 : 1-34, 2017
• 22 Russell, Stuart, "인공지능 1: 현대적 접근방식" 제이펍 1-, 2021
• 23 이광형, "인공지능" 한국방송통신대학교 출판문화원 2018
• 24 Sobel, Dava, "유리우주: 별과 우주를 사랑한 하버드 천문대 여성들" 사이언스걸스 2016
• 25 권보원, "양벌규정에 의한 기업 처벌은 실효적인 기업불법 통제수단인가: 통계분석으로 본 양벌규정 적용 현황" 서울대학교 대학원 2017
• 26 Goodfellow, Ian J., "심층 학습" 제이펍 2018
• 27 김관식, "신경망(Neural network) 기계학습에 기초한 인공지능 관련 발명의 법적 문제- 발명의 성립성과 발명자 적격성을 중심으로 -" 한국지식재산학회 (67) : 1-48, 2021
• 28 Holt, Nathalia, "로켓 걸스: 인간 컴퓨터라 불린 여인들" 사이언스걸스 2017
• 29 이긍희, "딥러닝의 통계적 이해" 한국방송통신대학교 출판문화원 2020
• 30 특허청, "기술분야별 심사실무가이드"
• 31 최경수, "국제지적재산권법" 한울아카데미 2017
• 32 Copeland, Jack, "계산하는 기계는 생각하는 기계가 될수 있을까?" 에디토리얼 2020
• 33 四宮和夫, "民法総則" 弘文堂 2018
• 34 Turing, Alan, "“계산 기계와 지능”, 지능에 관하여" HB프레스 2019
• 35 Kim, Daria, "‘AI-Generated Inventions’: Time to Get the Record Straight?" 69 : 443-, 2020
• 36 WIPO, "World IP Indicators 2021"
• 37 Moravec, Hans, "When Will Computer Hardware Match the Human Brain?" 1 : 10-, 1998
• 38 Grace, Katja, "When Will AI Exceed Human Performance? Evidence from AI Experts" 62 : 729-, 2018
• 39 Bakshi, Ashish, "What is Deep Learning? Getting Started with Deep Learning"
• 40 Bozinovski, Stevo, "Utjecaj slicnosti likova i transfera ucenja na obucavanje baznog perceptrona B2" 1976
• 41 Glorot, Xavier, "Understanding the Difficulty of Training Deep Feedforward Neural Networks" 2010
• 42 Hartmann, Christian, "Trends and Developments in Artificial Intelligence: Challenges to the Intellectual Property Rights Framework" European Commission 2020
• 43 Yalçin, Orhan G., "Top 5 Deep Learning Frameworks to Watch in 2021 and Why TensorFlow"
• 44 Vertinsky, Liza, "Thinking about Thinking Machines:Implications of Machine Inventors for Patent Law" 8 : 574-, 2002
• 45 Halevy, Alon, "The Unreasonable Effectiveness of Data" 24 : 8-, 2009
• 46 Polanyi, Michael, "The Tacit Dimension" Chicago University Press 1966
• 47 Chen, Hongming, "The Rise of Deep Learning in Drug Discovery" 23 : 1241-, 2018
• 48 Watson, David S., "The Rhetoric and Reality of Anthropomorphism in Artificial Intelligence" 29 : 417-, 2019
• 49 Thaler, Stephen L., "The Raman Spectrum of Neutron Irradiated Silicon" University of Missouri-Columbia 1982
• 50 Rosenblatt, Frank, "The Perceptron : A Probabilistic Model for Information Storage and Organization in the Brain" 65 : 386-, 1958
• 51 Hebb, Donald O, "The Organization of Behavior: A Neuropsychological Theory" John Wiley and Sons 1949
• 52 Blok, Peter, "The Inventor's New Tool: Artificial Intelligence - How Does It Fit in the European Patent System?" 39 : 69-, 2017
• 53 Kim, Won Joon, "The Innovation Landscape of Artificial Intelligence: The Global Competition to Build a Successful AI Ecosystem" KAIST ISPI & Clarivate 2021
• 54 Burkov, Andriy, "The Hundred-Page Machine Learning Book" 에이콘 2019
• 55 Smith, Gary, "The AI Delusion" Oxford University Press 2018
• 56 Kim, Daria, "Ten Assumptions about Artificial Intelligence That Can Mislead Patent Law Analysis" 2021
• 57 WIPO, "Technology Trends 2019: Artificial Intelligence" 2019
• 58 Straus, Joseph, "Studi di Diritto Commerciale per Vincenzo di Cataldo" Giappichelli 2021
• 59 Samuel, Arthur L., "Some Studies in Machine Learning Using the Game of Checkers" 3 : 210-, 1959
• 60 Polyak, Boris T., "Some Methods of Speeding Up the Convergence of Iteration Methods" 4 : 17-, 1964
• 61 Löwel, Siegrid, "Selection of Intrinsic Horizontal Connections in the Visual Cortex by Correlated Neuronal Activity" 255 : 209-, 1992
• 62 Ramachandran, Prajit, "Searching for Activation Functions" 2017
• 63 Banko, Michele, "Scaling to Very Very Large Corpora for Natural Language Disambiguation" 26 : 2001
• 64 Walch, Kathleen, "Rethinking Weak vs. Strong AI"
• 65 Delvaux, Mady, "Report with Recommendations to the Commission on Civil Law Rules on Robotics (2015/2103(INL))"
• 66 Newell, Allen, "Report on a General Problem Solving Program" 256 : 64-, 1959
• 67 Gramsci, Antonio, "Quaderni del Carcere, Volume primo" Giulio Einaudi 1977
• 68 Romm, Connor, "Putting the Person in PHOSITA : The Human’s Obvious Role in the Artificial Intelligence Era" 62 : 1413-, 2021
• 69 USPTO, "Public Views on Artificial Intelligence and Patent Policy"
• 70 Rosenblatt, Frank, "Principles of Neurodynamics: Perceptrons and the Theory of Brain Mechanisms" Spartan Books 1962
• 71 EPO, "Press Communiqué on decisions J 8/20 and J 9/20 of the Legal Board of Appeal"
• 72 Minsky, Marvin L., "Perceptrons: An Introduction to Computational Geometry" MIT Press 1969
• 73 Hopfield, John J., "Neural Networks and Physical Systems with Emergent Collective Computational Abilities" 79 : 2554-, 1982
• 74 Shapiro, Carl, "Navigating the Patent Thicket : Cross Licenses, Patent Pools, and Standard-Setting" 1 : 119-, 2001
• 75 Mitchell, Tom M., "Machine Learning" McGraw-Hill 1997
• 76 Srebrovic, Rob, "Leveraging the BERT algorithm for Patents with TensorFlow and BigQuery" Google 2020
• 77 Tieleman, Tijmen, "Lecture 6e-rmsprop: Divide the gradient by a running average of its recent magnitude" 2012
• 78 Rumelhart, David E., "Learning Representations by Back-propagating Errors" 323 : 533-, 1986
• 79 Rumelhart, David E., "Learning Internal Representations by Error Propagation" University of California San Diego Institute for Cognitive Science 1985
• 80 Brown, Tom B., "Language Models Are Few-Shot Learners"
• 81 Skansi, Sandro, "Introduction to Deep Learning: From Logical Calculus to Artificial Intelligence" Springer 2018
• 82 Hinton, Geoffrey E., "Improving Neural Networks by Preventing Co-adaptation of Feature Detectors" 2012
• 83 Krizhevsky, Alex, "ImageNet Classification with Deep Convolutional Neural Networks" 25 : 1097-, 2012
• 84 Abbott, Ryan, "I Think, Therefore I Invent: Creative Computers and the Future of Patent Law" 57 : 1079-, 2016
• 85 Jumper, John, "Highly Accurate Protein Structure Prediction with AlphaFold" 596 : 583-, 2021
• 86 Géron, Aurélien, "Hands-On Machine Learning with Scikit-Learn, Keras, and TensorFlow" 한빛미디어 2020
• 87 LeCun, Yann, "Gradient-Based Learning Applied to Document Recognition" 86 : 2278-, 1998
• 88 UKIPO, "Government Response to Call for Views on Artificial Intelligence and Intellectual Property"
• 89 Goodfellow, Ian J., "Generative Adversarial Nets" 27 : 2014
• 90 Zadeh, LotfiA., "Fuzzy Sets" 8 : 338-, 1965
• 91 Fabris, Daniele, "From the PHOSITA to the MOSITA : Will ‘Secondary Considerations’ Save Pharmaceutical Patents from Artificial Intelligence?" 51 : 685-, 2020
• 92 Azevedo, Frederico A. C., "Equal numbers of neuronal and nonneuronal cells make the human brain an isometrically scaled-up primate brain" 513 : 532-, 2009
• 93 Boyle, James, "Endowed by Their Creator?: The Future of Constitutional Personhood" 2011
• 94 EPO, "EPO publishes grounds for its decision to refuse two patent applications naming a machine as inventor"
• 95 윤성진, "Do it! 딥러닝 교과서" 이지스퍼블리싱 2021
• 96 Murray, Fiona, "Do Formal Intellectual Property Rights Hinder the Free Flow of Scientific Knowledge?: An Empirical Test of the Anti-Commons Hypothesis" 63 : 648-, 2007
• 97 He, Kaiming, "Delving Deep into Rectifiers: Surpassing Human-Level Performance on Imagenet Classification" 2015
• 98 Kamdar, Adi, "Defend Innovation: How to Fix Our Broken Patent System" 2015
• 99 Glorot, Xavier, "Deep Sparse Rectifier Neural Networks" 2011
• 100 Botvinick, Matthew, "Deep Reinforcement Learning and Its Neuroscientific Implications" 107 : 603-, 2020
• 101 Hale, Jeff, "Deep Learning Framework Power Scores 2018" 2018
• 102 Conlon, Ed, "DABUS: South Africa issues first-ever patent with AI inventor"
• 103 Moore, Gordon E., "Cramming More Components onto Integrated Circuits" 38 : 114-, 1965
• 104 Fraser, Erica, "Computers as Inventors-Legal and Policy Implications of Artificial Intelligence on Patent Law" 13 : 305-, 2016
• 105 Newell, Allen, "Computer Science as Empirical Inquiry:Symbols and Search" 19 : 113-, 1976
• 106 The AlphaCode team, "Competitive programming with AlphaCode"
• 107 Drexl, Josef, "Comments of the Max Planck Institute for Innovation and Competition of 11 February 2020 on the Draft Issues Paper of the World Intellectual Property Organization on Intellectual Property Policy and Artificial Intelligence" 2020
• 108 Heller, Michael A., "Can Patents Deter Innovation? The Anticommons in Biomedical Research" 280 : 698-, 1998
• 109 Allen, Tom, "Can Computer Make Contracts?" 9 : 25-, 1996
• 110 Werbos, Paul, "Beyond Regression: New Tools for Prediction and Analysis in the Behavioral Science" Harvard University 1974
• 111 Lillicrap, Timothy P., "Backpropagation and the Brain" 21 : 335-, 2020
• 112 WIPO, "Background Document on Patents and Emerging Technologies"
• 113 Feurer, Matthias, "Automated Machine Learning: Methods, Systems, Challenges" Springer 2019
• 114 Vaswani, Ashish, "Attention is All You Need" 2017
• 115 Aleksander, Igor, "Artificial Neuroconsciousness An Update" Springer 1995
• 116 Tong, Christopher, "Artificial Intelligence in Engineering Design 총 3권" Academic Press 1992
• 117 Mueller, John, "Artificial Intelligence for Dummies" John Wiley and Sons 2018
• 118 Chesterman, Simon, "Artificial Intelligence and the Limits of Legal Personality" 69 : 819-, 2020
• 119 Drexl, Josef, "Artificial Intelligence Systems as Inventors? A Position Statement of 7 September 2021 in View of the Evolving Case-Law Worldwide" Max Planck Institute for Innovation and Competition Research 2021
• 120 Center for the Fourth Industrial Revolution, "Artificial Intelligence Collides with Patent Law" 2018
• 121 Cybenko, George, "Approximation by Superpositions of a Sigmoidal Function" 2 : 303-, 1989
• 122 Hornik, Kurt, "Approximation Capabilities of Muitilayer Feedforward Networks" 4 : 251-, 1991
• 123 Hume, David, "An Enquiry Concerning Human Understanding" Oxford Univ. Press 2000
• 124 Senior, Andrew, "An Empirical Study of Learning Rates in Deep Neural Networks for Speech Recognition" 2013
• 125 Widrow, Bernard, "Adaptive Switching Circuits" 96 : 1960
• 126 Duchi, John, "Adaptive Subgradient Methods for Online Learning and Stochastic Optimization" 12 : 2011
• 127 Kingma, Diederik P., "Adam: A Method for Stochastic Optimization" 2014
• 128 岡本裕樹, "AIへの法人格付与に関する私法上の覚書(2·完)" (29) : 2020
• 129 岡本裕樹, "AIへの法人格付与に関する私法上の覚書(1)" (28) : 2020
• 130 Crevier, Daniel, "AI: The Tumultuous History of the Search for Artificial Intelligence" Basic Books 1993
• 131 배대헌, "AI 발명·특허출원을 통하여 살펴본 현행 특허법상 발명자 보호의 문제" 한국지식재산학회 (63) : 1-44, 2020
• 132 Ford, Martin, "AI 마인드: 세계적인 인공지능 개발자들이 알려주는 진실" 터닝포인트 2019
• 133 Hao, Karen, "AI pioneer GeoffHinton: ‘Deep learning is going to be able to do everything’" 2020
• 134 Floridi, Luciano, "AI and Its New Winter : from Myths to Realities" 33 : 1-, 2020
• 135 Burk, Dan L., "AI Patents and the Self-Assembling Machine" 105 : 301-, 2020
• 136 Dyson, Freeman, "A meeting with Enrico Fermi" 427 : 297-, 2004
• 137 McCarthy, John, "A Proposal for the Dartmouth Summer Research Project on Artificial Intelligence"
• 138 McCulloch, Warren S., "A Logical Calculus of the Ideas Immanent in Nervous Activity" 5 : 115-, 1943
• 139 Hinton, Geoffrey E., "A Fast Learning Algorithm for Deep Belief Nets" 18 : 1527-, 2006