“Fear, uncertainty and pain make people more susceptible to conspiracy theories”
We interviewed Dr Mariona Taulé i Delor, director of the Department of Catalan Philology and General Linguistics of the Faculty of Philology and Communication of the University of Barcelona (UB) and director of the research group of the Language and Computation Centre (CLiC) of the UB.
Her area of expertise is computational linguistics, which aims at the scientific study of speech and languages from a computational perspective, including both the development of computational language models and computer systems that allow language to be processed appropriately.
Dr Taulé received a Strategic Lines grant for a project that seeks to develop a strategy to counter infodemics, disinformation and, in particular, conspiracy theories in the healthcare area.
What is an infodemic?
An infodemic is an information pandemic, an epidemic of misinformation, such as for example the conspiracy theories that emerged during the Covid-19 pandemic. Infodemics are therefore a type of disinformation, that is, information that is deliberately false and designed to do harm. In fact, infodemic is a term used by the World Health Organisation (WHO) to refer to the flood of false and misleading messages related to Covid and vaccines.
How do new technologies affect the spread of false information?
The Internet and social media play an obvious role, as they are their main channel of dissemination. These types of messages mainly spread through social media, as fast or faster than a virus or an infectious disease, and perhaps they are also more difficult to combat. The information ecosystem has changed, and digital platforms facilitate the spread of these conspiracy theories, making it possible for them to reach a wider, more diverse audience very quickly. If add to that the pandemic and the resulting threat of a deadly disease, the lockdowns (less mobility, social isolation, more time available to access social media) and the uncertainty of science on how to deal with Covid, all this creates the ideal framework for mistrust of institutions as well as the need to understand what is happening, to find an explanation, which leads some people to turn to conspiracy theories.
Why?
Because these theories offer easy-to-understand explanations to confusing events when they claim that secret plots orchestrated by powerful people or malevolent groups are responsible for what is happening. All this makes the message more attractive and explains the growth of this type of disinformation and the interest it arouses. Fear, uncertainty, confusion, helplessness, loss of control, and pain make people more susceptible to believe in conspiracy theories, and that is why we study them.
You received a Strategic Lines grant for the project “XAI-DisInfodemics: EXplainable AI for disinformation and conspiracy detection during infodemics.” What is the current status of the project?
XAI-DisInfodemics is a project coordinated by the Polytechnic University of Valencia in collaboration with four other Spanish universities, the Spanish National Research Council (CSIC), and the company Symanto Spain SLU. Mireia Farrús, also a member of CLiC, and myself represent the University of Barcelona. It is a multidisciplinary team made up of psychologists, computational linguists and computer engineers.
The general objectives of the project are to develop a strategy to counter infodemics, disinformation and, in particular, conspiracy theories in the field of health; to analyse the elements that make conspiracy theories credible in social media; and to develop detection tools based on machine learning.
Our participation is focused on the creation and annotation of the digital corpora that will be used to create these detection tools, in other words, to train the different machine learning models that will be developed.
What is a digital corpus?
A digital corpus is a usually large collection of written or oral texts in electronic format that has been created for a specific purpose based on real examples of the use of a language or linguistic variety.
The corpus we are developing is multilingual, consisting of 5,000 messages in Spanish and 5,000 in English extracted from different Telegram channels suspected of spreading conspiracy theories. Annotation consists in differentiating messages that include ‘conspiracy thinking’ from those that show ‘critical thinking’, that is, to determine whether a message is critical of vaccines or of Covid-19 or of the way the pandemic has been handled but is not conspiratorial.
How can artificial intelligence help in the fight against disinformation?
It is very important to combat this type of disinformation – conspiracy theories – because they can cause real harm to people, their health and their physical safety, as UNESCO warned in 2020. Our fellow engineers involved in the project are in charge of developing the tools to detect this type of messages, and for this they need artificial intelligence, especially models based on deep learning techniques, and the corpora that we are developing. It is also necessary to understand how this information spreads through social media, which user profiles are more likely to share this type of messages and how we can identify them in order to create alert mechanisms and contain or slow down their propagation.
More about Dr Mariona Taulé
The best invention in history?
Writing, without it we would not be where we are. Writing fixes speech, which is ephemeral. Writing and reading open the door to civilisation. It allows us to preserve the memory of humanity and the accumulation of skills and knowledge.
What would you like to see in the future?
A society that includes diversity, a more equitable society, that guarantees equal opportunities, a fairer society. A society without hate messages, without false messages…. A utopia!
A future advancement that scares you?
The misuse that people may make of all the discoveries and technological advancements that are soon to be part of our lives.
A role model?
None in particular, but at the same time, many. In fact, all those people who have taught me, and will teach me something.
What can be done to achieve equality between men and women?
Rather than equality, I would rather speak of equity between men and women; we should be able to have the same opportunities, and this requires a lot of pedagogy. In my field of research, the best we can do is to develop language resources and automatic learning systems that are equitable and fair, without biased information, that are inclusive, that reflect diversity and multiple perspectives for understanding and interpreting the world.
The FBG is…
A fantastic team of people who support me when it comes to research management.