The advancements in the field of AI have brought about a broad range of benefits and have the potential to revolutionize many domains and industries. Large language models have recently produced a lot of buzz in artificial intelligence. These models are trained on enormous amounts of textual data and can generate meaningful text like humans. It can complete tasks such as translation and summarization and generate software codes. One of the key advantages of large language models is that they can easily generate good-quality text swiftly and at scale. This makes them beneficial for various applications, such as content creation, sentence completion, language translation, text summarization, etc.
The latest research conducted by Georgetown University, OpenAI, and Stanford Internet Observatory (SIO) is about how the expansion of the large language models can eventually lead to manipulating public viewpoints. The matter of concern is that the generative models, such as LLMs, that generate text can produce very convincible and unaltered data. It can be easily abused for trickery and negatively influencing people. If any spiteful individual uses the LLMs, it can promote a hoax agenda and be the reason for the escalation of fake news.
LLMs can generate human-like text, because of which they can be used to form fake news and impersonate real individuals or organizations. Moreover, since they can speedily generate large amounts of text, they can be used to overflow the internet with false information, making it harder for people to differentiate between what is true and what is wrong.
According to the research, a country’s citizens’ political influence on the web has grown in the past decade. Even The United States has gone through hundreds of cases of manipulation of public opinion. Be it a political event or a presidential election, the LLMs can mislead them all. Even the scale of the campaigns can be amplified with minimal costs making manipulation harder to detect.
The researchers used the ABC model of disinformation to examine how LLMs can be misused. The ABC model of disinformation is a framework that breakdown the various aspects that contribute to the escalation of false information. ‘A’ refers to ‘Actor,’ which can be a group of individuals who create and broadcast disinformation. ‘B’ stands for ‘Behaviour,’ which refers to the strategies used to spread propaganda. Lastly, ‘C’ stands for ‘Content’ which is untrue information. The team, through their research, mentioned that ABC are the three dimensions, and the actor can be real such as a politician who uses a bot or an automated technology to deceive people.
Large Language Models are not inherently malicious but have the potential to be wrongly used for manipulation and disinformation. Many studies are still pending to find a solution to this problem. Right now, careful attention should be paid to the type and source of news to avoid misuse, and users and developers should ethically use the model.
Check out the Report. All Credit For This Research Goes To the Researchers on This Project. Also, don’t forget to join our Reddit Page, Discord Channel, and Email Newsletter, where we share the latest AI research news, cool AI projects, and more.
Tanya Malhotra is a final year undergrad from the University of Petroleum & Energy Studies, Dehradun, pursuing BTech in Computer Science Engineering with a specialization in Artificial Intelligence and Machine Learning.
She is a Data Science enthusiast with good analytical and critical thinking, along with an ardent interest in acquiring new skills, leading groups, and managing work in an organized manner.