본문 바로가기
KERD
논문KCI 등재

금융에서의 생성형 인공지능 활용 현황과 법적 쟁점에 대한 연구

A Study on Current Utilization and Legal Issues of Generative Artificial Intelligence in Finance

이석준
한국증권법학회 (Korea Securities Law Association)|2024.04|증권법연구 (The Korean Journal of Securities Law)|제25권 제1호|pp. 65-102|DOI 10.17785/kjsl.2024.25.1.65

0

조회

0

다운로드

3

인용

0

참고문헌

키워드

생성형 인공지능Generative Artificial Intelligence챗봇편향Bias딥페이크금융Explainable AIFinanceChatbots생성적 적대 신경망트랜스포머 모델로보어드바이저설명가능한 인공지능자동화된 의사결정데이터 유출Deepfakes변형 자동 인코더Generative Adversarial Network(GAN)다중모드 인공지능Automated Decision MakingData LeakageRobo-advisorsTransformer ModelMulti-modal AI
주제분류법학

초록

GPT-3 출시 이래로 생성형 인공지능(이하 ‘AI’)은 전 세계에서 여러 산업 분야에 걸쳐 중요한 기능을 하고 있고 미래의 시장 규모도 빠르게 성장할 것으로 예측된다. 생성형 AI는 금융 분야에서 고객 상담, 고객 투자 관리, 대출 등 위험 관리, 금융회사의 업무 개선에 많은 기여를 하고 있다. 생성형 AI에는 생성적 적대 신경망(GAN), 트랜스포머 모델, 다중모드 AI 등이 있다. 위 AI들은 각각의 장․단점을 가지고 있으면서 합성 데이터의 생성, 사기성 거래 등의 구별, 비대면뱅킹, 금융흐름의 예측, 감정 분석, 재무 분석 등에 있어서 탁월한 성능을 발휘한다. 그럼에도 금융 분야에서 생성형 AI를 활용함에 있어 몇 가지 법적 문제들이 있을 것으로 예상된다. 여기에는 조작 위험(딥페이크), 금융 안정성의 위험(플래시크래쉬, 금융시스템 리스크), 투자자보호 관련 위험, 데이터 유출 위험, 불공정의 위험(데이터 및 알고리즘 편향, 설명가능한 인공지능, 자동화된 의사결정) 등이 있어 이에 초점을 두고 그 쟁점 및 법적․기술적 해결 방안에 대하여 살펴보았다. 나아가 고객 상담시 투명성 위험, 반경쟁 위험에 관하여도 짚어보았다.

Abstract

Since the release of GPT-3, generative artificial intelligence (hereinafter referred to as ‘AI’) has played a significant role across various industries worldwide, and the market size is predicted to grow substantially in the future. In the financial sector, generative AI is also bringing about positive changes, contributing significantly to tasks such as customer consultation, customer investment management, risk management including loans, and overall business improvement for financial companies. Generative AI includes Generative Adversarial Networks (GAN), Transformer models, and Multi-modal AI, each having its own strengths and weaknesses. These AIs excel in tasks such as generating synthetic data, distinguishing fraudulent transactions, non-face-to-face banking, predicting financial flows, sentiment analysis, financial analysis, and more. However, despite the positive impacts, there are anticipated legal issues in utilizing generative AI in the financial sector. These include manipulation risks (deepfakes), risks to financial stability (flash crashes, financial system risks), investor protection-related risks, data leakage risks, unfairness risks (data and algorithm bias, explainable AI, automated decision-making), among others. This discussion focuses on these points, examining the key issues and legal and technical solutions. Additionally, transparency risks during customer consultations and risks related to anti-competition are also touched upon.

참고문헌 (53)

  1. 이영철, "인공지능 알고리즘에 의한 불공정 거래행위의 법적 규제 - 공정거래법상 부당한 공동행위를 중심으로 -", 상사법연구, 통권 : 110 (2021)
  2. 주진열, "AI 알고리즘 가격설정과 이른바 ʻ알고리즘 묵시적 담합ʼ 문제에 대한 고찰", 경쟁법연구, 41 (2020), 한국경쟁법학회
  3. 2023 인공지능, "채명식", KISTEP 한국과학기술기획평가원, ~ (2023)
  4. 금융위원회, "금융분야 인공지능의 신뢰를 높인다", 보도자료, ~ (2023)
  5. 조선 Biz, "AI 챗봇 ‘이루다’, 개인정보유출 논란 속 결국 사실상 폐기", ~ (2021)
  6. ⑥ 김선우 우리은행 AI사업부장 “생성형 AI 활용한 PB 곧 나온다… 재무관리도 AI가 척척”, , 2023
  7. Alessio Azzutti, "Machine Learning, Market Manipulation, and Collusion on Capital Markets: Why the “Black Box” Matters", U. Pa.J. Int’l L, 43, pp.79~ (2021)
  8. Alexander Pomerantz, "Racial and Ethnic Differences in Insurer Classification of Nonemergent Pediatric Emergency Department Visits" (2023)
  9. Alžběta Krausová, "Intersections between Law and Artificial Intelligence", International Journal of Computer, 27(1) (2017)
  10. Andrew Kumiega, "Automated Finance : The Assumptions and Behavioral Aspects of Algorithmic Trading", Journal of Behavioral Finance, 13(1) (2012)
  11. Antonia Creswell, "Generative Adversarial Networks : An Overview", IEEE Signal Process Magazine, 35(1) (2018)
  12. Ashish Vaswani, "Attention Is All You Need", Neural Information Processing Systems (2017)
  13. The Judicial Demand for Explainable Artificial Intelligence, "Ashley Deeks", Virginia Public Law and Legal Theory Research, ~, pp.Virginia Public Law and Legal Theory Research Paper No. 2019-51 (2019)
  14. Binh Le, "Why Do Facial Deepfake Detectors Fail?", WDC '23: Proceedings of the 2nd Workshop on Security Implications of Deepfakes and Cheapfakes (2023)
  15. Caosen Xu, "A Financial Time-Series Prediction Model Based on Multiplex Attention and Linear Transformer Structure", Appl. Sci, 13(8) (2023)
  16. Dániel Kondor, "Towards matching user mobility traces in large-scale datasets", IEEE Transactions on Big Data, 6 (2020)
  17. Daniel McDuff, "Characterizing Bias in Classifiers using Generative Models", Advances in Neural Information Processing Systems, 32 (2019)
  18. David Gunning, "XAI—Explainable artificial intelligence", Science Robotics, 4(37) (2019)
  19. David S. Krause, "Mitigating Risks for Financial Firms Using Generative AI Tools", ~ (2023)
  20. Edmond Lezmi, "Time Series Forecasting with Transformer Models and Application to Asset Management", ~ (2023)
  21. Emilija Strelcenia, "A Survey on GAN Techniques for Data Augmentation to Address the Imbalanced Data Issues in Credit Card Fraud Detection", Mach. Learn. Knowl. Extr, 5(1) (2023)
  22. Enrique Martínez-Miranda, "Learning Unfair Trading: A Market Manipulation Analysis from the Reinforcement Learning Perspective", 2016 IEEE Conference on evolving and adaptive intelligent SYS (2016)
  23. Erik Brynjolfsson, "Generative AI at work", National bureau of economic research, ~ (2023)
  24. Government of Canada, "Guide on the use of Generative AI", ~
  25. Jinhong Wu, "Interpretation for Variational Autoencoder Used to Generate Financial Synthetic Tabular Data", Algorithms, 16(2) (2023)
  26. A Primer on Artificial Intelligence and Machine Learning for the Financial Services Industry, "Joerg Osterrieder", ~ (2023)
  27. Efficient Automation of Decision-making Processes in Financial Industry: Case Study and Generalised Model, "Juliana Hadjitchoneva", ~ (2019)
  28. Kelvin J.L. Koa, "Diffusion variational autoencoder for tackling stochasticity in multi-step regression stock price prediction", Proceedings of the 32nd ACM international conference on information and knowledge management (2023)
  29. Mahsa Tavakolia, "Multi-Modal Deep Learning for Credit Rating Prediction Using Text and Numerical Data Streams", ~ (2023)
  30. "Global Generative AI In Financial Services Market By Type (Solutions And Services), By Application (Credit Scoring, Fraud Detection, And Other), By Deployment Mode (Cloud And On-Premises), By Region And Companies - Industry Segment Outlook, Market Assessment, Competition Scenario, Trends, And Forecast 2023-2032", Market Researc report, ~ (2023)
  31. Max Schemmer, "Intelligent Decision Assistance Versus Automated Decision-Making: Enhancing Knowledge Work Through Explainable Artificial Intelligence", Proceedings of the 55th Hawaii International Conference on System Sciences (2022)
  32. National Secuirty Agency, "Contextualizing Deepfake Threats to Organizations", ~ (2023)
  33. Cloud and Threat Report: AI Apps in the Enterprise, "Netskope Threat Labs", ~ (2023)
  34. OECD, "Generative artificial intelligence in finance", ~ (2023)
  35. Rumelhart, "Learning representations by back-propagating errors", Nature, 323 (1986)
  37. Sasi Kumar Murakonda, "ML Privacy Meter: Aiding Regulatory Compliance by Quantifying the Privacy Risks of Machine Learning", CoRR, ~ (2020)
  38. Sepp Hochreiter, "Long Short-Term Memory", Neural Computation, 9 (1997)
  39. Siti Aishah Binti Mohd Yusof, "The Impact of Generative AI in Enhancing Credit Risk Modeling and Decision-Making in Banking Institutions", Emerging Trends in Machine Intelligence and Big Data, 15 (2023)
  40. Suman Kalia, "Potential impact of generative artificial intelligence(AI) on the financial industry", IJCI Journal, 12 (2023)
  41. Tanusree De, "Explainable AI: A Hybrid Approach to Generate Human-Interpretable Explanation for Deep Learning Prediction", Procedia Computer Science, 168 (2020)
  42. TC Helmus, "Artificial intelligence, deepfakes, and disinformation: A primer", Rand, ~ (2022)
  43. Tuna Tuncer, "Asset Price and Direction Prediction via Deep 2D Transformer and Convolutional Neural Networks", Proceedings of the Third ACM International Conference on AI in Finance (2022)
  44. Yu-Fu Chen, "Sentiment-influenced trading system based on multimodal deep reinforcement learning", Applied Soft Computing, 112(4) (2021)
  45. Ziyue Wang, "CNN‐and GAN‐based classification of malicious code families: A code visualization approach", International Journal of Intelligent Systems, 37 (2022)
  46. "How banks can adopt generative AI", American Banker, ~ (2023)
  47. "Google will label fake images created with its A.I.", CNBC, ~ (2023)
  48. "‘Verified’ Twitter accounts share fake image of ‘explosion’ near Pentagon, causing confusion", CNN, ~ (2023)
  49. "How Companies Can Use Generative AI And Maintain Data Privacy", Forbes, ~ (2023)
  50. "Nearly half of Indian internet users faced AI-driven voice scams this year", Techcircle, ~ (2023)
  51. "Voice Deepfakes Are Coming for Your Bank Balance", The New York Times, ~ (2023)
  52. "Fraudsters Used AI to Mimic CEO’s Voice in Unusual Cybercrime Case", The Wall Street Journal pro Cybersecurity, ~ (2019)
  53. 鹿島みかり, "投資判断におけるアルゴリズム․ AI の利用と法的責任: アルゴリズム․ AIの利用を巡る法律問題研究会報告書の概要", 金融研究 (2019), 日本銀行金融研究所

관련 논문AI추천

생성형AI의 은행업 적용에 있어서 법적 쟁점과 개선 방향구자영은행법학회 · 은행법연구2025 · 제18권 제2호생성형 인공지능 관련 범죄 위협 분류 및 대응 방안김민수, 박우빈, 박윤지, 유혜진, 정두원한국정보보호학회 · 정보보호학회논문지2024 · 제34권 제2호생성형 인공지능(Generative AI) 기술의 규제 방향에 대한 입법론적 고찰 - ChatGPT 등 인공지능 시스템 생성물에 대한 표시·고지의무를 중심으로 -김정화, 임동민, 차호동대검찰청 · 형사법의 신동향2023 · 제80호생성형 AI의 개발 및 이용에 관한 규제의 필요성 - 대규모 언어모델에 기반한 대화형 인공지능 서비스(LLMs AI)를 중심으로 -양 은 영법학연구원 · 성균관법학2023 · 제35권 제2호인공지능과 금융법김경태, 서완석법학연구소 · 가천법학2019 · 제12권 제4호생성형 AI 활용 단계에서의 개인정보 보호법적 쟁점과 기술적 해결방안육소영, 임을영, 현민성법학연구소 · The Journal of Law & IP2024 · 제14권 제2호

이 간행물의 다른 논문 (4건)

해외 주요국 탄소배출권 파생상품 감시체계 비교분석―EU, 미국, 우리나라를 중심으로―오석진pp. 1-64가상자산의 증권성 판단기준에 대한 연구이상근, 임승민, 양창규pp. 103-125가상자산․토큰증권의 담보권 설정 및 선의취득과 파산시 처리―가상자산을 물권의 객체로 인정하기 위한 방법에 관한 小考―오성헌, 신은혜, 전우정pp. 127-197신종 성과급여성 양도제한조건부 자기주식양도(RUS)의 法理 분석李哲松pp. 199-217
원문 보기