The rapid advancement of artificial intelligence has led to the emergence of deepfake, digital media that has been manipulated to replace a person's likeness with another. This technology has seen significant improvements, becoming easier to use and producing results increasingly difficult to distinguish from reality. This development has raised ethical discussions surrounding its deceiving nature. Furthermore, deepfakes have had a considerable impact and application on social media, enabling their spread. Despite this, the public discourse on social media, along with its societal and personal values associated with deepfakes, remains underexplored. This study addresses this gap by examining social media discourse and perception surrounding the prominent ethical concerns of deepfakes, and situating these concerns within the broader landscape of AI ethics. Through a qualitative method resembling netnography, 320 posts from Reddit and Youtube were thematically analyzed through a passive observation, along with their respective comment section. The findings reveal various concerns, surrounding misinformation and consent to deeper fears about deepfakes' role in fostering distrust, as well as more abstract apprehensions regarding the technology's abuse and harmful applications. These concerns further revealed how generalized established AI ethical principles might be interpreted in the deepfake context, also showing how and why these principles might be violated by this technology. Particularly it revealed terms how principles such as dignity, transparency, privacy and non-maleficence might be diverged in deepfake applications.
| Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:uu-523078 |
| Date | January 2023 |
| Creators | Abdul Hussein, Mohamed, Bogren, William |
| Publisher | Uppsala universitet, Informationssystem |
| Source Sets | DiVA Archive at Upsalla University |
| Language | English |
| Detected Language | English |
| Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
| Format | application/pdf |
| Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0019 seconds