Could AI be a Solution against the Far Right?
The Role of ChatGPT in Mitigating the Risks of Extremism
In recent years, the growth of extremist groups has been an escalating concern across the globe. Technology, often seen as an enabler of these ideologies, particularly through its potential to produce fake news and deep fakes, can also play a crucial role in combating them. An emerging tool with significant potential in this context is ChatGPT. This article explores how ChatGPT can help mitigate the risks associated with the rise of extremist groups, focusing on two main aspects: its role in combating scientific denialism and its ability to promote moderation and civility in social interactions.
Combating Scientific Denialism and Promoting Moderate Stances
ChatGPT, along with other LLMs (large language models) like Claude or Gemini, is designed to provide politically correct and academically high-quality responses. In some tasks, these AI tools perform comparably to well-educated humans, whose brains are trained to achieve good results in exams and competitions like the US Bar Exam, which qualifies law graduates to practice. They process vast amounts of information and generate responses that reflect a broad knowledge base and an ability to produce coherent answers adhering to the grammatical rules of cultivated language. This capability to provide relatively accurate and useful information, although not always correct or representative of our society’s cultural diversity, makes ChatGPT a powerful tool against scientific denialism.
Those who have used ChatGPT or its competitors in recent months know that although it is a probability-based text generator, it surprisingly simulates a conversation with a well-informed individual who is familiar with classic works of literature, art, philosophy, science, religion, psychoanalysis; it knows geography, history, can solve mathematical problems better than many, and can even give mental health advice, though always recommending consultation with a real expert, thereby valuing human expertise which it cannot replace.
The machine lacks the ability to perceive subtleties such as irony in natural language processing; resolve ambiguities; understand emotion in medical care; genuinely create and innovate rather than merely producing variations of existing concepts; and make ethical decisions with the same flexibility as humans, to name a few examples. If you ask ChatGPT if it can replace human professionals, it will confirm this. However, keep in mind that neither I nor ChatGPT necessarily make true statements. To be more precise, I believe that no one can guarantee the truth of their statements, but the scientific method teaches us to verify the information received until it is accepted as evidence. Apparently, ChatGPT is fed with scientific evidence and a selection of the best content humanity has produced. It remains to be seen how this content is curated and what algorithm is used to process these data. We do not have enough information to discuss this matter yet. All I can say is that I use ChatGPT extensively and have been very impressed and satisfied with the results.
In times where misinformation spreads rapidly through social networks, the presence of a relatively reliable agent that consistently responds based on scientific evidence can be a weapon against the spread of conspiracy theories and false narratives. The hype around LLM technology like ChatGPT, coupled with its accessibility, attracts a broad audience that is exposed to information reinforcing the importance of the scientific method and verifiable data. Although LLMs are not designed to verify the data they produce, or to use the scientific method themselves, they generate content highly likely to be aligned with established science.
Moreover, ChatGPT’s stance tends to be moderate and balanced. This is particularly relevant in the current polarized scenario, where extremist discourses find fertile ground. ChatGPT’s moderation serves as an example of balanced political positioning, offering an alternative to the dystopian views that often dominate online discussions. By providing answers that avoid extremes and promote rational dialogue, ChatGPT contributes to a healthier and more constructive discussion environment.
Promoting Cordiality and Breaking Relationship Bubbles
Another key aspect of ChatGPT is its ability to interact in a friendly, patient, and open manner with anyone. This trait is particularly important in a context where social interactions are often marked by polarization and conflict, frequently exacerbated by cancel culture. ChatGPT, with its consistently cordial approach, provides an example of how interactions can be constructive and respectful, even amidst disagreements.
ChatGPT’s ability to break through relationship bubbles is crucial. Social networks tend to create bubbles where people are exposed only to information and opinions that reinforce their pre-existing beliefs. This lack of exposure to divergent perspectives can lead to a type of radicalization that tends to impoverish one’s understanding of the world. ChatGPT, by being available to interact with a wide variety of individuals, can introduce moderate viewpoints and encourage dialogue among different groups. This mediating role can promote mutual understanding.
Furthermore, by offering an example of constant moderation and cordiality, ChatGPT can positively influence human interactions, even though this machine behavior may not be ideal for a person with a healthy brain, who can benefit from the varied emotions they are capable of producing. At a time when cancel culture threatens open dialogue, the presence of an agent that encourages patience and empathy can help restore damaged relationships, including among family members. This positive influence on social interactions can, in the long term, contribute to a more cohesive and resilient society against the influence of extremist ideologies, reducing the level of harmful intolerance so prevalent in our society today.
The first draft of this text was written in Portuguese by ChatGPT 4-o under the guidance and review of Bernardo Monteiro. Translation to english by ChatGPT 4.0.
