Article from World Economic Forum
Disinformation is a “pandemic” that has engulfed economies large and small, and is being described as a threat to the stability of society. Around the world, people are facing threats to their lives and personal safety due to the abundance of misinformation, which manipulates people's perceptions or propagates lies. This is causing emotional distress and social disruption.
Artificial Intelligence (AI)-based programs are being used to create ""deepfakes" of political leaders by adjusting videos, audio, and photos. Deepfakes can be used to create chaos in society. Artificial intelligence is also getting better at generating human-like content using language models like GPT-3 that can author articles, poems, and essays. The forgery of all types of content has become so common with AI that open-source software like face swap and DeepFaceLab can allow even the most discreet amateurs to be at the epicenter of social disharmony. At a time when people don't know where to trust anymore, "technology for good" seems to be the only savior.
Semantic analytics for basic filtering
To combat disinformation with technology is content analysis. AI-based tools can perform linguistic analysis of textual content and detect cues, including word patterns, syntax, and readability, to distinguish computer-generated content from human-generated text. Such algorithms can take any piece of text and check for word vectors, word placement, and connotation to identify traces of hate speech. In addition, AI algorithms can reverse engineer manipulated images and videos to detect deepfakes and highlight content that should be reported.
But that’s not enough: adversarial networks are becoming so sophisticated that algorithms will soon produce content that is indistinguishable from that created by humans. Algorithms with such semantic analysis cannot interpret the content within hate speech images that have not been manipulated, but instead are distributed with the wrong context, malicious intent, or additional content. They also cannot check whether the claims made by some pieces of content are false. Language barriers also add to the challenges. Essentially, the sentiment of an online post can be assessed, but not its veracity. This is where human intervention with AI is required.
Root tracing: next-level cop
Fake news is often found to have the same root – the place of origin before the news spread. The project Fandango, for example, uses stories that have been flagged as false by human fact-checkers and then searches for social media posts or websites that have similar words or claims. This allows journalists and experts to trace fake stories to their roots and eliminate any potential threats before they spread out of control.
Services such as Politifact, Snopes and FactCheck employ human editors who can perform the initial research required to verify the authenticity of a report or image. Once a fake is found, AI algorithms help search the web and combat similar pieces of content that could fuel social discord. If the content turns out to be genuine, the website's article can be assigned a reputation score. Trust Project uses parameters such as sources, references, ethical standards, and corrections to assess the credibility of news outlets.
Spreading tests to stop the spread
There is a noticeable difference between the way fake news and real news circulate on social media. Researchers from MIT suggest that fake news circulates six times faster than real news to reach 1500 people on Twitter.
People usually share fake news more quickly without much critical thought or a form of judgment. GoodNews uses an AI engine to identify fake news using engagement metrics, as fake news shows more shares than likes, compared to real news. Such techniques to catch suspicious content based on its spread could help prevent radicalization.
People in essence
Using technology is a reactionary step when the world needs a proactive approach to combat disinformation. Artificial intelligence will be successful if we educate the masses – especially young people – to be vigilant against misinformation. Fake news is not simply a matter of algorithms, but of the philosophy behind how we handle knowledge – good or bad. Communities of informed users can contribute to ethical monitoring activities, while collaborative knowledge gathering between professional organizations is essential for verifying raw information.
Humanizing the approach to combating disinformation should be the highest priority in order to build a well-informed society of critical thinkers. The lack of proactive measures that involve all actors can lead to the rapid erosion of trust in the media and institutions, which is a precursor to anarchy. Until people learn to objectively evaluate online content, AI-based technologies should be our ally in combating online disinformation.