Deepfake

IA for discrimination and generation of digital content

Authors

  • Thaïs Helena Falcão Botelho Pontifical Catholic University of São Paulo, Faculty of Exact Sciences and Technology, Postgraduate Program in Intelligence Technologies and Digital Design, São Paulo, São Paulo, Brazil.
  • Winfried Nöth Pontifícia Universidade Católica de São Paulo, Faculdade de Ciências Exatas e Tecnologia, Programa de Pós-Graduação em Tecnologias da Inteligência e Design Digital, São Paulo, São Paulo, Brasil.

DOI:

https://doi.org/10.23925/1984-3585.2021i23p69-78

Keywords:

Deepfake, Fake news, Artificial Intelligence, GANs, Education, Media literacy

Abstract

The term “fake news” has become popular in the social media in 2016 during the elections for the presidency of the United States. Studies have shown that false news may have a greater number of likes than messages on reputable sites. They may even influence the outcome of elections. The production of false news uses technological resources from the world of print media, photos, texts, and layout. Currently, the digital media predominate. In addition to the printed text, they use audiovisual content. In this virtual environment, new artificial intelligence technologies are being developed, as is the case of deepfake, which can also be used for the production of content for disseminating false news. Such placements can threaten trust in institutions and democracy. One of the paths proposed to combat deepfakes is educational work through media literacy.

Author Biographies

Thaïs Helena Falcão Botelho, Pontifical Catholic University of São Paulo, Faculty of Exact Sciences and Technology, Postgraduate Program in Intelligence Technologies and Digital Design, São Paulo, São Paulo, Brazil.

PhD student and Master in Technologies of Intelligence and Digital Design, PUC - SP. Member of the Sociotramas group, PUC-SP. Editor and researcher of images for educational materials.

Winfried Nöth, Pontifícia Universidade Católica de São Paulo, Faculdade de Ciências Exatas e Tecnologia, Programa de Pós-Graduação em Tecnologias da Inteligência e Design Digital, São Paulo, São Paulo, Brasil.

Professor at the Graduate Program in Intelligence Technologies and Digital Design (TIDD/PUC-SP).

References

COMITÊ GESTOR DA INTERNET NO BRASIL. TIC domicílios 2019: principais resultados. São Paulo: CGI, 2020. Disponível em: cetic.br/media/analises/tic_domicilios_2019_coletiva_imprensa.pdf. Acesso em: 10 abr. 2021.

DACK, Sean. Deep fakes, fake news, and what comes next. The Henry M. Jackson School of International Studies, University of Washington. Washington, 20 mar. 2019. Disponível em: jsis.washington.edu/news/deep-fakes-fake-news-and-what-comes-next/. Acesso em: dez. 2020.

DATA SCIENCE ACADEMY. Deep learning book. São Paulo: Data Science Academy, 2021. Disponível em: deeplearningbook.com.br. Acesso em: abr. 2021.

GOODFELLOW, Ian J. et al. Generative adversarial networks. arXiv preprint arXiv:1406.2661, 2014. Disponível em: arxiv.org/abs/1406.2661. Acesso em: abr. 2021

GUNTHER, Richard; NISBET, Erik C.; BECK, Paul. Trump may owe his 2016 victory to “fake news”, suggests a new study. The Conversation, 15/02/2018. Disponível em: theconversation.com/trump-may-owe-his-2016-victory-to-fake-news-new-study-suggests-91538. Acesso em: 1 abr. 2021.

JONES, Katie. How COVID-19 has impacted media consumption, by generation. Visual Capitalist, Vancouver, 7 abr. 2020. Disponível em: visualcapitalist.com/media-consumption-covid-19. Acesso em: abr. 2021.

KINAST, Priscilla. Os incríveis números do Youtube em 2019: quantos vídeos tem no Youtube? Qual vídeo mais assistido? Quantas pessoas usam o Youtube? Quais são os maiores canais? Quantas horas são assistidas a cada minuto no Youtube? Essas e outras respostas aqui. Oficina da Net, 7 ago. 2019. Disponível em: oficinadanet.com.br/tecnologia/26607-os-incriveis-numeros-do-youtube-em-2019. Acesso em: 30 mar. 2021

PARK, Sung-Wook; HUH, Jun-Ho; KIM, Jong-Chan. BEGAN v3: avoiding mode collapse in GANs using variational inference. Electronics, 9(4): 688, 2020. Disponível em: doi.org/10.3390/electronics9040688. Acesso em: 30 mar. 2021.

PRESS, Larry. A real-names domain registration policy would discourage political lying. CircleID, 17, nov. 2016. Disponível em: circleid.com/posts/ 20161117_real_names_domain_registration_policy_discourage_political_lying. Acesso em: 30 mar. 2021.

SILVERMAN, Craig. This analysis shows how viral fake election news stories outperformed real news on Facebook. BuzzFeed.News. 16 nov. 2016. Disponível em: buzzfeednews.com/article/craigsilverman/viral-fakeelection-news-outperformed-real-news-on-facebook#.uc9gevywE. Acesso em: 1 abr. 2021.

TIME 100: The 100 most influential people. Time, 21 abril, 2016. Disponível em: time.com/collection/2016-time-100/leaders/. Acesso em: 1 abr. 2021.

WESTERLUND, Mika. The emergence of deepfake technology: a review. Technology Innovation Management Review, v. 9, n. 11, 2019. Disponível em: timreview.ca/article/1282. Acesso em: 30 mar. 2021.

YUGE, Cláudio. YouTube vai usar mais IA e menos revisão humana em conteúdos sobre coronavírus. Canaltech, 16 mar. 2020. Disponível em: canaltech.com.br/internet/youtube-vai-usar-mais-sua-ia-e-pode-removermais-conteudo-sobre-coronavirus-161927/. Acesso em: 1 abr. 2021.

Published

2021-10-12

Issue

Section

Artigos

Most read articles by the same author(s)

1 2 > >>