As Good As A Coin Toss: Human detection of AI-generated images, videos, audio, and audiovisual stimuli
2403.16760
0
0
🔎
Abstract
As synthetic media becomes progressively more realistic and barriers to using it continue to lower, the technology has been increasingly utilized for malicious purposes, from financial fraud to nonconsensual pornography. Today, the principal defense against being misled by synthetic media relies on the ability of the human observer to visually and auditorily discern between real and fake. However, it remains unclear just how vulnerable people actually are to deceptive synthetic media in the course of their day to day lives. We conducted a perceptual study with 1276 participants to assess how accurate people were at distinguishing synthetic images, audio only, video only, and audiovisual stimuli from authentic. To reflect the circumstances under which people would likely encounter synthetic media in the wild, testing conditions and stimuli emulated a typical online platform, while all synthetic media used in the survey was sourced from publicly accessible generative AI technology. We find that overall, participants struggled to meaningfully discern between synthetic and authentic content. We also find that detection performance worsens when the stimuli contains synthetic content as compared to authentic content, images featuring human faces as compared to non face objects, a single modality as compared to multimodal stimuli, mixed authenticity as compared to being fully synthetic for audiovisual stimuli, and features foreign languages as compared to languages the observer is fluent in. Finally, we also find that prior knowledge of synthetic media does not meaningfully impact their detection performance. Collectively, these results indicate that people are highly susceptible to being tricked by synthetic media in their daily lives and that human perceptual detection capabilities can no longer be relied upon as an effective counterdefense.
Create account to get full access
Overview
- As synthetic media becomes more realistic and accessible, it is increasingly being used for malicious purposes like fraud and nonconsensual pornography.
- Traditionally, the main defense against being misled by synthetic media has been the ability of humans to visually and auditorily discern real from fake.
- This study assessed how accurately people can distinguish synthetic images, audio, video, and audiovisual content from authentic content.
Plain English Explanation
This research looked at how good people are at telling the difference between real and fake digital content, like images, audio, and videos. As AI-generated media gets more advanced and easier to create, it's being used for harmful things like financial fraud and nonconsensual pornography.
Traditionally, the way to spot fake content has been for people to use their senses - looking and listening carefully to figure out what's real. But the researchers wanted to see how well this actually works in practice. They had over 1,200 people try to tell the difference between real and AI-generated images, audio, videos, and a mix of audio and video.
The results showed that people really struggle to reliably detect synthetic content, even when it's just a single image or audio clip. Their ability to spot fakes gets even worse when the content has multiple elements like both audio and video. The researchers also found that people do worse at detecting fakes if the content is in a language they don't know well.
Overall, this suggests that relying on people's senses to catch fake digital content isn't a reliable defense anymore, now that the technology to create realistic fakes is so advanced and accessible.
Technical Explanation
The researchers conducted a perceptual study with 1,276 participants to assess how accurately people could distinguish synthetic media from authentic content. They tested participants' ability to detect synthetic images, audio-only, video-only, and audiovisual stimuli sourced from publicly available generative AI technology.
The testing conditions and stimuli were designed to emulate typical online platforms where people might encounter synthetic media in their daily lives. The researchers found that overall, participants struggled to meaningfully discern between synthetic and authentic content. Detection performance worsened when:
- The stimuli contained synthetic content vs. authentic content
- The images featured human faces vs. non-face objects
- The stimuli was a single modality (image, audio, or video) vs. multimodal (audio and video)
- The audiovisual stimuli had mixed authenticity vs. being fully synthetic
- The content was in a foreign language vs. a language the observer was fluent in
Importantly, the researchers also found that prior knowledge of synthetic media did not significantly impact participants' detection performance. These results indicate that people are highly susceptible to being deceived by synthetic media in their everyday lives, and that human perceptual capabilities can no longer be relied upon as an effective defense against such deception.
Critical Analysis
The study provides valuable insights into the growing threat of synthetic media, but it also has some limitations. While the researchers aimed to reflect real-world conditions, the experiment was still conducted in a controlled setting. It's possible that people's ability to detect fakes could be even worse in more naturalistic situations where they are not specifically focused on the task.
Additionally, the study did not explore how factors like personal biases, emotional responses, or contextual cues might influence people's perception of synthetic media. These elements could play a significant role in how individuals interpret and react to fake content in their daily lives.
Further research is needed to better understand the various psychological and social factors that contribute to the vulnerability of humans to synthetic media. Longitudinal studies, cross-cultural comparisons, and investigations into the potential countermeasures and interventions could provide a more comprehensive understanding of this complex issue.
Conclusion
This study's findings suggest that as synthetic media becomes increasingly sophisticated and widespread, people's ability to reliably distinguish real from fake content is severely limited. This has significant implications for how we navigate the digital landscape, as we can no longer depend on our senses alone to protect us from being misled or deceived.
Addressing this challenge will require a multifaceted approach, involving technological solutions, educational initiatives, and policy changes. Continued research and vigilance will be crucial as we work to stay ahead of the rapidly evolving synthetic media landscape and its potential for harm.
This summary was produced with help from an AI and may contain inaccuracies - check out the links to read the original source documents!
Related Papers
Unmasking Illusions: Understanding Human Perception of Audiovisual Deepfakes
Ammarah Hashmi, Sahibzada Adil Shahzad, Chia-Wen Lin, Yu Tsao, Hsin-Min Wang
0
0
The emergence of contemporary deepfakes has attracted significant attention in machine learning research, as artificial intelligence (AI) generated synthetic media increases the incidence of misinterpretation and is difficult to distinguish from genuine content. Currently, machine learning techniques have been extensively studied for automatically detecting deepfakes. However, human perception has been less explored. Malicious deepfakes could ultimately cause public and social problems. Can we humans correctly perceive the authenticity of the content of the videos we watch? The answer is obviously uncertain; therefore, this paper aims to evaluate the human ability to discern deepfake videos through a subjective study. We present our findings by comparing human observers to five state-ofthe-art audiovisual deepfake detection models. To this end, we used gamification concepts to provide 110 participants (55 native English speakers and 55 non-native English speakers) with a webbased platform where they could access a series of 40 videos (20 real and 20 fake) to determine their authenticity. Each participant performed the experiment twice with the same 40 videos in different random orders. The videos are manually selected from the FakeAVCeleb dataset. We found that all AI models performed better than humans when evaluated on the same 40 videos. The study also reveals that while deception is not impossible, humans tend to overestimate their detection capabilities. Our experimental results may help benchmark human versus machine performance, advance forensics analysis, and enable adaptive countermeasures.
5/8/2024
Harnessing Machine Learning for Discerning AI-Generated Synthetic Images
Yuyang Wang, Yizhi Hao, Amando Xu Cong
0
0
In the realm of digital media, the advent of AI-generated synthetic images has introduced significant challenges in distinguishing between real and fabricated visual content. These images, often indistinguishable from authentic ones, pose a threat to the credibility of digital media, with potential implications for disinformation and fraud. Our research addresses this challenge by employing machine learning techniques to discern between AI-generated and genuine images. Central to our approach is the CIFAKE dataset, a comprehensive collection of images labeled as Real and Fake. We refine and adapt advanced deep learning architectures like ResNet, VGGNet, and DenseNet, utilizing transfer learning to enhance their precision in identifying synthetic images. We also compare these with a baseline model comprising a vanilla Support Vector Machine (SVM) and a custom Convolutional Neural Network (CNN). The experimental results were significant, demonstrating that our optimized deep learning models outperform traditional methods, with DenseNet achieving an accuracy of 97.74%. Our application study contributes by applying and optimizing these advanced models for synthetic image detection, conducting a comparative analysis using various metrics, and demonstrating their superior capability in identifying AI-generated images over traditional machine learning techniques. This research not only advances the field of digital media integrity but also sets a foundation for future explorations into the ethical and technical dimensions of AI-generated content in digital media.
5/27/2024
🔍
How to Distinguish AI-Generated Images from Authentic Photographs
Negar Kamali, Karyn Nakamura, Angelos Chatzimparmpas, Jessica Hullman, Matthew Groh
0
0
The high level of photorealism in state-of-the-art diffusion models like Midjourney, Stable Diffusion, and Firefly makes it difficult for untrained humans to distinguish between real photographs and AI-generated images. To address this problem, we designed a guide to help readers develop a more critical eye toward identifying artifacts, inconsistencies, and implausibilities that often appear in AI-generated images. The guide is organized into five categories of artifacts and implausibilities: anatomical, stylistic, functional, violations of physics, and sociocultural. For this guide, we generated 138 images with diffusion models, curated 9 images from social media, and curated 42 real photographs. These images showcase the kinds of cues that prompt suspicion towards the possibility an image is AI-generated and why it is often difficult to draw conclusions about an image's provenance without any context beyond the pixels in an image. Human-perceptible artifacts are not always present in AI-generated images, but this guide reveals artifacts and implausibilities that often emerge. By drawing attention to these kinds of artifacts and implausibilities, we aim to better equip people to distinguish AI-generated images from real photographs in the future.
6/14/2024
🔗
Finding AI-Generated Faces in the Wild
Gonzalo J. Aniano Porcile, Jack Gindi, Shivansh Mundra, James R. Verbus, Hany Farid
0
0
AI-based image generation has continued to rapidly improve, producing increasingly more realistic images with fewer obvious visual flaws. AI-generated images are being used to create fake online profiles which in turn are being used for spam, fraud, and disinformation campaigns. As the general problem of detecting any type of manipulated or synthesized content is receiving increasing attention, here we focus on a more narrow task of distinguishing a real face from an AI-generated face. This is particularly applicable when tackling inauthentic online accounts with a fake user profile photo. We show that by focusing on only faces, a more resilient and general-purpose artifact can be detected that allows for the detection of AI-generated faces from a variety of GAN- and diffusion-based synthesis engines, and across image resolutions (as low as 128 x 128 pixels) and qualities.
4/8/2024