Generative AI as 'potent weapon and shield' in battle of political misinformation

By Haziq Jeelani* 
In the pulsating heart of the digital era, the political arena is ceaselessly molded by the swift and relentless flow of information. The line between fact and fiction often blurs, creating a nebulous landscape where truth and deceit intertwine. 

The advent of generative artificial intelligence (AI) has added a new dimension to this dynamic, wielding the power to both fuel and quell the spread of misinformation.

Pandora's box of generative AI

Generative AI, with its uncanny ability to craft realistic text, images, and videos, holds the potential to significantly amplify the problem of misinformation in politics. Deepfake technology, a subset of generative AI, can fabricate eerily convincing videos of politicians uttering words they never spoke or performing actions they never took, sowing seeds of discord and confusion among voters. 

Similarly, AI algorithms can generate counterfeit news articles or social media posts, meticulously tailored to specific audiences, potentially swaying public opinion and influencing the outcome of elections.
In the 2020 Gabonese presidential election, an AI-generated video of President Ali Bongo was circulated, in which he appeared to be in robust health despite having suffered a stroke. The video was later unmasked as a deepfake, raising alarm bells about the use of AI to manipulate public opinion during political campaigns. 

In the 2016 U.S. presidential election, AI was reportedly used to generate propaganda and counterfeit news stories that were then disseminated on social media platforms. These stories were designed to stoke political divisions and influence voter behavior.
AI-powered bots have been deployed to spread misinformation on social media platforms. For instance, during the Brexit referendum in the UK, automated bots were used to amplify certain viewpoints, creating an illusion of widespread support or opposition to Brexit. 

AI can also be used to analyze vast amounts of data on individuals' online behavior, enabling political campaigns to micro-target specific groups with tailored misinformation. This was a significant concern during the Cambridge Analytica scandal, where the data of millions of Facebook users was used without consent to influence voter behavior in the 2016 U.S. presidential election.

Shield of machine learning against misinformation

On the flip side, machine learning, another branch of AI, can be a potent tool in the fight against misinformation. Advanced algorithms can be trained to detect and flag potential misinformation, helping to stem its spread.

Data of millions of Facebook users was used to influence voter behavior in the 2016 U.S. presidential election

For instance, a recent study titled "Va ex ccine sentiment analysis using BERT + NBSVM and geo-spatial approaches" leverages a combination of BERT (Bidirectional Encoder Representations from Transformers) and NBSVM (Naive Bayes and Support Vector Machine) for sentiment analysis. This approach identifies and classifies people's feelings about vaccines, thereby providing useful insights to counter misinformation.

Several advanced algorithms and systems can be employed to prevent the spread of misinformation. Machine learning models can be trained on large datasets to recognize patterns associated with misinformation, such as certain phrasing or source reliability. Natural Language Processing (NLP) techniques can be used to analyze the content of news articles or social media posts, identifying potential misinformation based on the language used. 

Moreover, the use of semi-supervised learning, as discussed in "A systematic literature review of cyber-security data repositories and performance assessment metrics for semi-supervised learning," can be beneficial in situations where only a few labels are necessary or available for building robust models.
In conclusion, while generative AI poses a significant challenge in the fight against misinformation in politics, machine learning offers a promising solution. By leveraging advanced algorithms and systems, it is possible to detect and counteract misinformation, ensuring that the political discourse remains grounded in fact rather than fiction.
---
*PhD student at Claremont Grad (USA), consults on AI. Twitter: @hazytalks