The World of Generative AI: Deepfakes and Large Language Models
Alakananda Mitra, Saraju P. Mohanty, Elias Kougianos
TL;DR
This paper analyzes the interplay between deepfakes and large language models (LLMs) within Generative AI, focusing on societal risks such as misinformation and manipulation of democratic processes. It surveys what deepfakes are and how they are created using autoencoders and generative adversarial networks, then surveys the evolution and capabilities of transformer-based LLMs and multimodal variants. It highlights how ChatGPT and other chatbots can streamline deepfake production by generating dialogue and scripts, enabling scalable, low-cost synthetic media through platforms and services. Finally, it reviews current governance and defense efforts, including safety benchmarks, platform policies, and proposals like digital watermarking and provenance tracking, while calling for stronger regulation and real-time detection technologies to mitigate risks in the near term.
Abstract
We live in the era of Generative Artificial Intelligence (GenAI). Deepfakes and Large Language Models (LLMs) are two examples of GenAI. Deepfakes, in particular, pose an alarming threat to society as they are capable of spreading misinformation and changing the truth. LLMs are powerful language models that generate general-purpose language. However due to its generative aspect, it can also be a risk for people if used with ill intentions. The ethical use of these technologies is a big concern. This short article tries to find out the interrelationship between them.