Affiliation:
1. Bennett University, India
2. University of Technology and Applied Sciences, Ibri, Oman
3. University of Naples, Italy
Abstract
As generative AI advances, deepfakes are proliferating in sophistication and accessibility, spurring an arms race between media synthesis and detection. This chapter traces the evolution of deepfakes, focusing on algorithms like GANs, VAEs in enhancing realism, and predicts future trajectories, including hyper-realistic media, streamlined creation, and widespread benign and malicious adoption. Despite constructive applications in entertainment, education, marketing and medicine, threats loom regarding misinformation, consent violations, and propagating social biases. The authors emphasize the need for comprehensive solutions through public awareness campaigns, advanced digital forensics, ethical legal frameworks, incentivizing “blue sky” innovation, and social media oversight. Navigating societal implications requires collective vigilance and forward-looking perspectives. This chapter underscores the importance of proactive, reasoned preparation as increasingly disruptive technologies emerge.
Reference27 articles.
1. Ajder, H., Patrini, G., Cavalli, F., & Cullen, L. (2019). The state of deepfakes: Landscape, threats, and impact. Amsterdam: Deeptrace, 27.
2. Deepfakes Generation and Detection: A Short Survey
3. A Survey on Variational Autoencoders from a Green AI Perspective
4. Brock, A., Donahue, J., & Simonyan, K. (2018). Large scale GAN training for high fidelity natural image synthesis. ArXiv Preprint ArXiv:1809.11096.
5. Ten years of generative adversarial nets (GANs): a survey of the state-of-the-art