EUVIMEDEUVIMED entdecken
Crossref

Deepfakes and Political Misinformation in U.S. Elections

Tom Sorell

Techné: Research in Philosophy and Technology · 2023 · 27 (3) · 363-386

DOI: 10.5840/techne20231110190 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Techné: Research in Philosophy and Technology, 2023. Verfasst von Tom Sorell. Die Arbeit wurde bisher 4-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Audio and video footage produced with the help of AI can show politicians doing discreditable things that they have not actually done. This is deepfaked material. Deepfakes are sometimes claimed to have special powers to harm the people depicted and their audiences—powers that more traditional forms of faked imagery and sound footage lack. According to some philosophers, deepfakes are particularly “believable,” and widely available technology will soon make deepfakes proliferate. I first give reasons why deepfake technology is not particularly well suited to producing “believable” political misinformation in a sense to be defined. Next, I challenge claims from Don Fallis and Regina Rini about the consequences of the wide availability of deepfakes. My argument is not that deepfakes are harmless, but that their power to do major harm is highly conditional in liberal party political environments that contain sophisticated mass-media.

Zitieren

APA

Sorell, T. (2023). Deepfakes and Political Misinformation in U.S. Elections. Techné: Research in Philosophy and Technology, 27(3), 363-386. https://doi.org/10.5840/techne20231110190

Vancouver

Sorell T. Deepfakes and Political Misinformation in U.S. Elections. Techné: Research in Philosophy and Technology. 2023;27(3):363-386. doi:10.5840/techne20231110190

BibTeX
@article{Sorell2023,
  author = {Sorell, Tom},
  title = {Deepfakes and Political Misinformation in U.S. Elections},
  journal = {Techné: Research in Philosophy and Technology},
  year = {2023},
  volume = {27},
  number = {3},
  pages = {363--386},
  doi = {10.5840/techne20231110190}
}
Kostenlos starten →