A data-driven analysis of how AI-driven misinformation and deepfakes affect public trust in US financial institutions

Abstract

The rapid development of artificial intelligence has fundamentally changed the digital communication environment, allowing spreading information more quickly, as well as creating highly realistic and deceptive content. Misinformation and deepfakes with the involvement of AI has become one of the greatest threats to civil perception, institutional validity, and social stability. Financial ecosystems, in which trust, transparency, and accurate information are the most important factors, spreading deceptive or manipulated content may provoke confusion in the population, distort the perception of risks, and even the loss of legitimacy of the U.S. financial institutions. This study provides a quantitative analysis of these threats in the form of large-scale fact-checking content of the Verified Posts: Fact-Checking Online Content (Politifact) data set. The sample used consists of over 20,000 online posts which have been labeled within the various credibility categories, such as true, false, half-true, pants-fire and mostly-true, covering the years 2008 to 2022. This study through the methods of Natural Language Processing and machine-learning algorithms explores linguistic frames, topic trends, misinformation rates, and propaganda time series. Temporal analysis indicates that the wave of misinformation is related to significant socio-political developments, which indicates that there are external forces, and the amplification of the algorithm. Even though the dataset consists of no video deepfakes, the text-based patterns of misinformation patterns are pretty similar to the manipulation tactics used in the production of synthetic media. This study shows that false accounts, particularly those who concern policy choices, economic outcomes, or governmental activities, can indirectly affect the trust of the population in financial institutions as they change the views of their institutional competence, justness, and stability. The present research makes a contribution to a better comprehension of AI-facilitated misinformation ecosystem evolution and the effect they can have on financial trust. The lessons derived can provide a practical policy to regulators and financial institutions and policy makers to improve the digital governance, enhance the level of awareness and create a strong mitigation policy of misinformation. To promote credibility and sustainability of the financial system in the long term, it is critical to strengthen information integrity so that the society can develop a sense of confidence in the system.