In the digital era, the spread of online disinformation has become a pervasive issue, influencing public opinion, shaping political landscapes, and even affecting global events. The year 2024 stands out as a critical juncture in this ongoing battle, marked by significant advancements in artificial intelligence (AI) and its application in identifying and countering fake news. This blog post delves into the evolution of AI in combating online disinformation, in the context of 2024 being the year when elections will be held in most countries around the world, engaging 2 Billion voters.
The Past: The Onset of AI in Disinformation Detection
Reflecting on the past, the 2016 US Presidential Elections serve as a pivotal moment in the history of online disinformation. This period was characterized by manually created disinformation content, which was then amplified across social media platforms by bots. The content primarily consisted of written material, supplemented by images such as memes and altered pictures. The year 2016 coined the term fake news and made it mainstream.
During this time, the primary defense against disinformation included human fact-checkers, digital literacy programs, and automated verification systems. However, these measures struggled to scale with the sheer volume of disinformation and mostly had a post-event impact, limiting their effectiveness in real-time situations.
The Present: AI at the Forefront of Disinformation Combat
Fast forward to 2023, the landscape has evolved dramatically with the emergence of Large Language Models (LLMs) and generative AI, leading to the automatic creation of multi-modal content. This era marked significant advances in AI-generated texts, images, and even videos, raising the bar for disinformation tactics.
However, with these advancements came new challenges. The “uncanny valley” phenomenon is not relevant anymore (where AI-generated content could be distinguished from human-generated content) which poses a substantial threat to information integrity, as multi-modal LLMs and Generative AI tools that can create high-quality human-like videos: deepfakes. The internet became inundated with unsupervised AI-generated content, including social media posts, product descriptions, and user reviews, blurring the lines between authentic and fabricated information.
In response to these challenges, AI technologies like TrustServista took center stage. These platforms utilized deep learning approaches such as Recurrent Neural Networks (RNNs) and Transformers to collect data, analyze it, and provide alerts. They offered a range of services, including named entity recognition, automatic classification, text extraction from images (OCR), sentiment analysis, topic clustering, and stance detection, thus equipping users with comprehensive tools to identify and counter disinformation.
A notable innovation in this era was the development of image similarity detection tools like SimSearchNet++ employed by Facebook which significantly enhanced the ability to detect altered images, a common tactic in disinformation campaigns.
The Future: AI as the Ultimate Defender of Truth
As we step into 2024, the biggest global election year in history, the question that looms large is: “Will Democracy survive it?” The answer lies in the continuous development and deployment of AI algorithms capable of detecting AI-generated content.
The key question for this year is the crucial role of AI: will it aid, or combat disinformation? With content being produced en masse, human verification becomes impractical, necessitating the development of AI algorithms that can operate at internet scale. These algorithms are not only expected to detect AI-generated content with high accuracy but also ensure that fundamental human rights are safeguarded in the process.
In conclusion, the journey from 2016 to 2024 highlights a significant transformation in the battle against online disinformation. AI, once a budding technology, has now become an indispensable tool in identifying and countering fake news. As the complexity of disinformation tactics evolves, so does the sophistication of AI solutions, promising a future where truth and authenticity prevail in the digital realm. The continuous innovation and adaptation of AI technologies remain our best hope in safeguarding democracy and ensuring the integrity of information in the years to come.