An official website of the United States government
The proliferation of inflammatory or misleading “fake” news content has become increasingly common in recent years. Simultaneously, it has become easier than ever to use AI tools to generate photorealistic images depicting any scene imaginable. Combining these two—AI-generated fake news content—is particularly potent and dangerous. To combat the spread of AI-generated fake news, we propose the MiRAGeNews Dataset, a dataset of 12,500 high-quality real and AI-generated image-caption pairs from state-of-the-art generators. We find that our dataset poses a significant challenge to humans (60% F-1) and state-of-the-art multi-modal LLMs (< 24% F-1). Using our dataset we train a multi-modal detector (MiRAGe) that improves by +5.1% F-1 over state-of-the-art baselines on image-caption pairs from out-of-domain image generators and news publishers. We release our code and data to aid future work on detecting AI-generated content.
more »
« less
Defending Against Neural Fake News
Recent progress in natural language generation has raised dual-use concerns. While applications like summarization and translation are positive, the underlying technology also might enable adversaries to generate neural fake news: targeted propaganda that closely mimics the style of real news. Modern computer security relies on careful threat modeling: identifying potential threats and vulnerabilities from an adversary's point of view, and exploring potential mitigations to these threats. Likewise, developing robust defenses against neural fake news requires us first to carefully investigate and characterize the risks of these models. We thus present a model for controllable text generation called Grover. Given a headline like `Link Found Between Vaccines and Autism,' Grover can generate the rest of the article; humans find these generations to be more trustworthy than human-written disinformation. Developing robust verification techniques against generators like Grover is critical. We find that best current discriminators can classify neural fake news from real, human-written, news with 73% accuracy, assuming access to a moderate level of training data. Counterintuitively, the best defense against Grover turns out to be Grover itself, with 92% accuracy, demonstrating the importance of public release of strong generators. We investigate these results further, showing that exposure bias -- and sampling strategies that alleviate its effects -- both leave artifacts that similar discriminators can pick up on. We conclude by discussing ethical issues regarding the technology, and plan to release Grover publicly, helping pave the way for better detection of neural fake news.
more »
« less
- Award ID(s):
- 1714566
- PAR ID:
- 10107992
- Date Published:
- Journal Name:
- Neurips
- Format(s):
- Medium: X
- Sponsoring Org:
- National Science Foundation
More Like this
-
-
Automated journalism technology is transforming news production and changing how audiences perceive the news. As automated text-generation models advance, it is important to understand how readers perceive human-written and machine-generated content. This study used OpenAI’s GPT-2 text-generation model (May 2019 release) and articles from news organizations across the political spectrum to study participants’ reactions to human- and machine-generated articles. As participants read the articles, we collected their facial expression and galvanic skin response (GSR) data together with self-reported perceptions of article source and content credibility. We also asked participants to identify their political affinity and assess the articles’ political tone to gain insight into the relationship between political leaning and article perception. Our results indicate that the May 2019 release of OpenAI’s GPT-2 model generated articles that were misidentified as written by a human close to half the time, while human-written articles were identified correctly as written by a human about 70 percent of the time.more » « less
-
null (Ed.)In this paper, we examine the long-neglected yet important effects of point sampling patterns in point cloud GANs. Through extensive experiments, we show that sampling-insensitive discriminators (e.g.PointNet-Max) produce shape point clouds with point clustering artifacts while sampling-oversensitive discriminators (e.g. PointNet++, DGCNN) fail to guide valid shape generation. We propose the concept of sampling spectrum to depict the different sampling sensitivities of discriminators. We further study how different evaluation metrics weigh the sampling pattern against the geometry and propose several perceptual metrics forming a sampling spectrum of metrics. Guided by the proposed sampling spectrum, we discover a middle-point sampling-aware baseline discriminator, PointNet-Mix, which improves all existing point cloud generators by a large margin on sampling-related metrics. We point out that, though recent research has been focused on the generator design, the main bottleneck of point cloud GAN actually lies in the discriminator design. Our work provides both suggestions and tools for building future discriminators. We will release the code to facilitate future research.more » « less
-
Machine Learning-Based Identifications of COVID-19 Fake News Using Biomedical Information ExtractionThe spread of fake news related to COVID-19 is an infodemic that leads to a public health crisis. Therefore, detecting fake news is crucial for an effective management of the COVID-19 pandemic response. Studies have shown that machine learning models can detect COVID-19 fake news based on the content of news articles. However, the use of biomedical information, which is often featured in COVID-19 news, has not been explored in the development of these models. We present a novel approach for predicting COVID-19 fake news by leveraging biomedical information extraction (BioIE) in combination with machine learning models. We analyzed 1164 COVID-19 news articles and used advanced BioIE algorithms to extract 158 novel features. These features were then used to train 15 machine learning classifiers to predict COVID-19 fake news. Among the 15 classifiers, the random forest model achieved the best performance with an area under the ROC curve (AUC) of 0.882, which is 12.36% to 31.05% higher compared to models trained on traditional features. Furthermore, incorporating BioIE-based features improved the performance of a state-of-the-art multi-modality model (AUC 0.914 vs. 0.887). Our study suggests that incorporating biomedical information into fake news detection models improves their performance, and thus could be a valuable tool in the fight against the COVID-19 infodemic.more » « less
-
Abstract The efficacy of fake news corrections in improving memory and belief accuracy may depend on how often adults see false information before it is corrected. Two experiments tested the competing predictions that repeating fake news before corrections will either impair or improve memory and belief accuracy. These experiments also examined whether fake news exposure effects would differ for younger and older adults due to age-related differences in the recollection of contextual details. Younger and older adults read real and fake news headlines that appeared once or thrice. Next, they identified fake news corrections among real news headlines. Later, recognition and cued recall tests assessed memory for real news, fake news, if corrections occurred, and beliefs in retrieved details. Repeating fake news increased detection and remembering of corrections, correct real news retrieval, and erroneous fake news retrieval. No age differences emerged for detection of corrections, but younger adults remembered corrections better than older adults. At test, correct fake news retrieval for earlier-detected corrections was associated with better real news retrieval. This benefit did not differ between age groups in recognition but was greater for younger than older adults in cued recall. When detected corrections were not remembered at test, repeated fake news increased memory errors. Overall, both age groups believed correctly retrieved real news more than erroneously retrieved fake news to a similar degree. These findings suggest that fake news repetition effects on subsequent memory accuracy depended on age differences in recollection-based retrieval of fake news and that it was corrected.more » « less
- Free Publicly Accessible Full Text
-
Accepted Manuscript1.0
- Conference Paper:
- The DOI is not currently available.
