-Sandeep Khurana, Research Fellow, IIDS

Online misinformation and disinformation have engulfed nations globally. Diverse, multiple legal regimes globally make enforcement difficult as the digital world cuts across national boundaries seamlessly, but the law does not.

Introduction

Online misinformation and disinformation have rapidly engulfed nations globally over the last decade, posing a severe threat to democracies. Sadly, social media, intended to empower the voiceless, has yet to last that initial hope and promise. Despite early flourish (Arab Spring, BP Oil spill, etc.) and sporadic positive usage in challenging hegemonic and autocratic forces through connective action, instant communication, and information processing potential of social media, the counterforces have ensured that dominant talk today is of threats and hazards of social media. Political resources and potential for political gains have converged with commercial motives of platforms to ensure social goals are not just secondary but easily sacrificed and compromised.

What ails the research in misinformation and disinformation?

That Multi-disciplinary research is complicated and requires tremendous effort is apparent. However, there is more. A research domain dealing with political power requires the researcher's political, social, and moral courage to unravel and speak the unadulterated truth. Often, the researched truth is contended with the most potent counterforces, and quite expectedly, the sponsors are either absent or unwilling. Worse, research incentives could be perverse or distracting, missing the wood for the trees. The relatively dynamic and novel landscape is another factor that discourages research on the subject. Mass, online, and social spaces are recent, and we borrow our understanding from the physical town square even though the two are different. Paradigm has shifted and is still evolving.

While one would consider social media a public space, a publicly owned public commons, the all-critical data is not public and remains proprietary. Algorithms remain opaque. Most research is based on available data rather than data collected by asking pertinent questions about the research subject.

As with research, the law field is also grappling with accelerated ushering into the world of technology and data. Thus, the demand for instant decisions and justice, even as legal jurisprudence, is stuck in the colonial era. This has implications for research. Given human creativity, both freedom of speech and hate speech are nuanced. A balance between hate and liberty is a tough ask. An Innuendo, a dog whistle, body language, even a mischievous pause, a punctuation mark, or a strategically placed emoji could easily alter an innocent tweet to a new meaning. Legal precedence can be inconsistent, and good laws take long to stabilize. Another legal and regulatory dilemma is anonymity, which helps unearth the truth against autocratic or corrupt regimes. Still, it can unleash toxicity without repercussions for the anonymous user. Law is a double-edged sword when gathering evidence in digital trials that can be used as evidence and the counter-use of technology to get around the same. The viral spread of information also poses a challenge as the scale of harm is high. Still, the reaction time available is much less, even negligible, than the routine time required to mobilize resources. Diverse, multiple legal regimes globally can also make enforcement difficult as the digital world cuts across national boundaries seamlessly, but the law does not.

The way modern research is focused on business, not social outcomes, makes it difficult to get sponsors for areas like these. Applied research with high real-world value is not even the goal of most conferences and journals.

Research Directions and Contributions to Policy

One of the most popular research areas is analyzing content for toxicity, polarization, or topic modeling for various themes. Newer algorithms using NLP techniques of all hues are added daily to the global repertoire. Other ML tools like predictive analytics and recommender engines are helpful for businesses and hence attract interest, too. Marketers have an apparent interest in influencer analysis. Few users exercise disproportionate influence (power law), making them the subject of interest.  

A challenge is the use of closed group messaging and communication platforms. Striking a balance between privacy and the needs of law enforcement makes hate sneak through without the scrutiny of researchers' analyses. Obtaining data creatively from such spaces is also a prospect for research.

Beyond these popular research topics on business, technical, and engineering in social media, there is interest from behavioral sciences, too. Public policy, legal research, psychology, sociology, and political science also research widespread online hate speech. Economists and game theorists look to alter incentive structures in online interactions to influence and disincentivize misinformation.

Studies of information diffusion, social network analysis, and static and dynamic networks provide insights into structural frameworks of information and misinformation spread that are not readily available from conventional tools. Low-cost tools that fake news and propaganda use have resulted in unscrupulous players in domains like healthcare, finance, and business using these, too. Consequently, research in many fields now contends with the subject.  

Legal AI to tackle misinformation and disinformation is a promising area, and applied NLP, specialized legal ontologies, and custom frameworks offer analytics opportunities to collaborate with legal experts. The landscape is vast, and the listing in this section is indicative and needs to be more comprehensive. With new techniques to solve these challenges and accompanying news of more unique ways to use deep fakes or misuse ChatGPT in a disruptively dystopian way, the area is not static in time and is evolving dynamically.

Endnote:

The need of the hour is for data scientists to step up to contribute to actionable, applied research and collaborate across other domains, particularly with policy-makers, to solve these burning problems of society.

1Author has published ongoing research work in social movements on social media, hate speech, and social network analysis. His other research areas are e-commerce and board interlock.