알고리즘과 메타포:페이스북의 ‘에코챔버’와 구글 검색의 ‘필터버블’을 중심으로 = Algorithm and Metaphor: On Facebook’s ‘Echo Chamber’ and Google Search’s ‘Filter Bubble’

다국어 초록 (Multilingual Abstract)

Algorithms used by large online platforms such as Facebook are often pointed out as a major driving force of political polarization. Algorithms, especially with their personalization features, are alleged to narrow the scope of thinking and fuel extremism by offering each user tailored and biased content. However, algorithms were not regarded as a formidable force from the beginning. This paper argues that two metaphors, ‘echo chamber’ and ‘filter bubble’, contributed to the demarcation of the algorithm from the Internet’s complex ecosystem. Echo chamber refers to a closed cyberspace where similarly minded people amplify each other’s voices, and filter bubble points to an isolated world created by personalization. These two metaphors gained popularity in the 2010s, especially after Brexit and Trump’s win, and led researchers to investigate the Internet’s one-sided influence upon the unknowing public. Even though these experiments largely failed to bring to light the existence of echo chambers or filter bubbles, they served as boundary-making practices that isolated the algorithm as an autonomous and dangerous actor. In the process, the meaning of echo chamber was transformed to place algorithms as the core of its mechanism. Indeed, as scholars of science studies have long observed, metaphors go beyond simply describing objects as they are, inciting novel research programs and actively shaping scientific theories and objects.
This paper finishes by presenting four arguments against portraying algorithms as sovereign power and subsequently vilifying it. Firstly, algorithms cannot operate by themselves and necessarily cooperate with humans in reciprocal systems. Secondly, most problems of extremism and discrimination generated by artificial intelligence are not caused by algorithms per se but rather the data incorporated. Thirdly, people do not blindly accept every piece of information fed to them by the system. Lastly, most of the content shared on the Internet is intertwined with consumer demand generated both online and offline.

국문 초록 (Abstract)

페이스북과 같은 대형 온라인 플랫폼의 알고리즘은 정치 양극화를 심화하는 요인으로 자주 지목된다. 특히 알고리즘의 개인화(personalization) 기능은 각 사용자의 습관과 취향에 따라 콘텐츠를 선별적으로 제공함으로써, 사고의 폭을 좁히고 사회의 분열을 심화한다고 평가받는다. 그러나 알고리즘이 처음부터 위협적인 행위자였던 것은 아니다. 본 논문은 알고리즘이 인터넷의 복잡한 생태계로부터 분리되는 데에 ‘에코챔버’(echo chamber)와 ‘필터버블’(filter bubble) 메타포가 기여했음을 보인다. 에코챔버는 비슷한 사람끼리 모여 서로의 견해를 증폭시키는 현상, 필터버블은 개인이 맞춤화된 세계에 갇혀 편협한 시각을 기르는 현상을 가리킨다. 두 메타포는 2010년대, 특히 브렉시트와 트럼프 대통령 당선 직후 많은 인기를 끌며, 컴퓨터공학자와 언론학자들 사이에서 페이스북과 구글 검색 각각을 중심으로 에코챔버와 필터버블의 실체를 포착하려는 여러 실험을 유발하였다. 이 실험들은 의도한 경험적 근거는 찾지 못했지만, 부차적인 효과로서 알고리즘을 독립되고 위험한 존재자로 경계 지었다. 이 과정에서 에코챔버의 정의는 초기와 달리 알고리즘의 독자적인 행위성을 강조하는 식으로 변형되기도 했다. 이는 과학기술 분야에서 메타포가 그저 주어진 대상을 묘사하는 도구가 아니라, 과학 이론과 대상을 예상치 못한 형태로 구체화하는 힘임을 보여준다.
글의 결론에서는 알고리즘을 자율적이고 불가항력의 힘으로 보는 인식의 문제점을 네 가지 지적한다. 첫째로, 알고리즘은 사람에게 일방적으로 작용하지 않고, 필연적으로 사람과 상호의존적인 관계에 있다. 둘째로, 인공지능이 발생시키는 혐오와 차별의 문제 대부분은 알고리즘 자체보다는 그와 결합된 데이터에서 비롯한다. 셋째로, 사람은 알고리즘이 무엇을 보여주든지 의심 없이 수용하는 수동적인 존재가 아니다. 마지막으로, 온라인으로 널리 소비되는 많은 콘텐츠는 인터넷 안팎에서 형성된 수요와 얽혀 있다.

참고문헌 (Reference)

1 Coons, C., 2021

2 하대청, "주의 통치하기: 유튜브 추천시스템과 심리-알고리즘적 주체" 한국과학기술학회 22 (22): 145-172, 2022

3 고학수 ; 박도현 ; 이나래, "인공지능 윤리규범과 규제 거버넌스의 현황과 과제" 서울대학교 법학연구소 13 (13): 7-36, 2020

4 엘리 프레이저, "온라인 "필터 버블"을 주의하세요"

5 데이비드 섬프터, "알고리즘이 지배한다는 착각" 해나무 2022

6 엘리 프레이저, "생각 조종자들" 알키 2011

7 아네마리 몰, "바디 멀티플" 그린비 2022

8 채진원, "무엇이 우리 정치를 위협하는가" 인물과사상사 2016

9 캐시 오닐, "대량살상 수학무기" 흐름출판 2017

10 김동인, "[특집] 신문 대신 유튜브 보고, ‘성향 같아야 신뢰’"

11 Leetaru, K., "Why 2017 was the Year of the Filter Bubble?"

12 Kay, L., "Who Wrote the Book of Life? : A History of the Genetic Code" Stanford University Press 2000

13 Oremus, W., "Who Controls Your Facebook Feed"

14 Simonite, T., "When It Comes to Gorillas, Google Photos Remains Blind"

15 Facebook, "What ‘The Social Dilemma’ Gets Wrong"

16 Sandvig, C., "What Facebook’s ‘It’s Not Our Fault’ Study Really Means"

17 Krafft, T. D., "What Did You See? A Study to Measure Personalization in Google’s Search Engine" 8 (8): 1-23, 2019

18 D’Onfro, J., "We Sat in on an Internal Google Meeting Where They Talked About Changing the Search Algorithm — Here’s What We Learned"

19 Google SearchLiaison, "Twitter"

20 Thune, J., "Thune, Colleagues Reintroduce Bipartisan Bill to Increase Internet Platform Transparency"

21 Lewis, R., "This is What the News Won’t Show You" : YouTube Creators and the Reactionary Politics of Micro-celebrity" 21 (21): 201-217, 2020

22 Orlowski, J., "The Social Dilemma"

23 Wright, T., "The Platform Transparency and Accountability Act: New legislation addresses platform data secrecy" Stanford Cyber Policy Center

24 Mittelstadt, B., "The Ethics of Algorithms : Mapping the Debate" 3 (3): 1-21, 2016

25 Cinelli, M., "The Echo Chamber Effect on Social Media" 118 (118): 1-8, 2021

26 Levin, S. A., "The Dynamics of Political Polarization" 118 (118): 2116950118-, 2021

27 Zuiderveen Borgesius, F. J., "Should We Worry About Filter Bubbles?" 5 (5): 1-16, 2016

28 Munger, K., "Right-Wing YouTube : A Supply and Demand Perspective" 27 (27): 186-219, 2022

29 Sunstein, C. R., "Republic.com" Princeton University Press 2001

30 Obama, B., "President Obama’s Farewell Address"

31 Pew Research Center, "Political Polarization in the American Public"

32 Arendt, F., "Optimizing Online Suicide Prevention : A Search Engine-Based Tailored Approach" 32 (32): 1403-1408, 2017

33 Walker, M., "News Consumption Across Social Media in 2021" Pew Research Center

34 Smith, R. E., "My Social Media Feeds Look Different from Yours and it’s Driving Political Polarization"

35 Black, M., "Models and Metaphors : Studies in Language and Philosophy" Cornell University Press 1962

36 Boyd, R., "Metaphor and Thought" Cambridge University Press 481-532, 1979

37 Barad, K., "Meeting the Universe Halfway: Quantum Physics and the Entanglement of Matter and Meaning" Duke University Press 2006

38 Gillespie, T., "Media Technologies: Essays on Communication, Materiality, and Society" MIT Press 167-193, 2014

39 DuckDuckGo, "Measuring the "Filter Bubble": How Google is Influencing What You Click"

40 Hannak, A., "Measuring Personalization of Web Search" 527-538, 2013

41 Price, R., "Mark Zuckerberg Denies That Facebook is Trapping its Users in ‘Filter Bubbles’" Insider

42 Rahwan, I., "Machine Behavior" 568 : 477-486, 2019

43 Bono, J. J., "Literature and Science: Theory and Practice" Northeastern University Press 59-90, 1990

44 Bruns, A., "It’s Not the Technology, Stupid: How the ‘Echo Chamber’ and ‘Filter Bubble’ Metaphors Have Failed Us" International Association for Media and Communication Research 1-12, 2019

45 Flaxman, S., "Ideological Segregation and the Effects of Social Media on News Consumption" 80 : 1-41, 2013

46 Meta, "How machine learning powers Facebook’s News Feed ranking algorithm"

47 Google, "How Results are Automatically Generated"

48 Meyers, P. J., "How Often Does Google Update its Algorithm?"

49 Tufekci, Z., "How Facebook’s Algorithm Suppresses Content Diversity (Modestly) and How the Newsfeed Rules Your Clicks"

50 DuckDuckGo, "Help Spread Privacy"

51 Meta, "Hard Questions: Is Spending Time on Social Media Bad for Us?"

52 Ziewitz, M., "Governing Algorithms : Myth, Mess, and Methods" 41 (41): 3-16, 2016

53 Metz, C., "Google Is 2 Billion Lines of Code—And It’s All in One Place"

54 Delaney, K., "Filter Bubbles are a Serious Problem with News, Says Bill Gates"

55 Flaxman, S., "Filter Bubble, Echo Chambers, and Online News Consumption" 80 (80): 298-320, 2016

56 Hagey, K., "Facebook Tried to Make Its Platform a Healthier Place. It Got Angrier Instead"

57 Bail, C. A., "Exposure to Opposing Views on Social Media Can Increase Political Polarization" 115 (115): 9216-9221, 2018

58 Bakshy, E., "Exposure to Ideologically Diverse News and Opinion on Facebook" 348 (348): 1130-1132, 2015

59 Hempel, J., "Eli Pariser Predicted the Future. Now He Can’t Escape It"

60 Garrett, K., "Echo Chambers Online?: Politically Motivated Selective Exposure Among Internet News Users" 14 : 265-285, 2009

61 Sunstein, C. R., "Echo Chamber: Bush v. Gore, Impeachment, and Beyond" Princeton University Press 2001

62 오요한 ; 홍성욱, "Does Artificial Intelligence Algorithm Discriminate Certain Groups of Humans?" 한국과학기술학회 18 (18): 153-215, 2018

63 Passi, S., "Data Vision: Learning to See Through Algorithmic Abstraction"

64 Gunning, D., "DARPA’s Explainable AI(XAI)Program : A Retrospective" 2 (2): e61-, 2021

65 Olson, K. K., "Cyberspace as Place and the Limits of Metaphor" 11 (11): 10-18, 2005

66 The YouTube Team, "Continuing our work to improve recommendations on YouTube"

67 Cutts, M., "Comment on “DuckDuckGo: Escape Your Search Engine Filter Bubble"

68 Courtois C., "Challenging Google Search Filter Bubbles in Social and Political Information : Disconforming Evidence From a Digital Methods Case Study" 35 : 2006-2015, 2018

69 Castelvecchi, D., "Can We Open the Black Box of AI?" 538 (538): 20-23, 2016

70 Thomas, S. L., "Algorithms as Fetish: Faith and Possibility in Algorithmic Work" 1-11, 2018

71 Metz, R., "Algorithms are Everywhere. Here’s Why You Should Care"

72 Haim, M., "Abyss or Shelter? On the Relevance of Web Search Engines’ Search Results When People Google for Suicide" 32 (32): 253-258, 2017

73 Bineham, J., "A Historical Account of the Hypodermic Model in Mass Communication" 55 (55): 230-246, 2009

74 Dahlgren, P. M., "A Critical Review of Filter Bubbles and a Comparison with Selective Exposure" 42 (42): 15-33, 2021

75 Sunstein, C. R., "#Republic: Divided Democracy in the Age of Social Media" Princeton University Press 2017