Un journalisme de qualité peut vaincre le contenu généré par l'IA

Pamela Duncan, a former journalist for the British newspaper The Guardian who currently lectures at University College Dublin in Ireland, argued during a recent television interview that high-quality journalism retains the capacity to outperform content generated by artificial intelligence. Speaking on HRT’s program Razgovor s razlogom, Duncan presented findings from research conducted with colleagues regarding how open Facebook groups contribute to societal radicalization in the United Kingdom and facilitate the spread of far-right ideologies. Her analysis focused specifically on the aftermath of the anti-immigration riots that erupted in 2024 following a stabbing incident in Southport. The catalyst for these disturbances was an attack by Axel Rudakubana, who was seventeen years old at the time. He entered a children’s dance workshop and killed three girls with a knife while injuring several others. The unrest quickly expanded beyond Southport to various other regions across the UK. Duncan stated that her team sought to understand the individuals involved in these riots, including those who spread rumors online or incited further disorder. To achieve this, the researchers examined court reports to gather available data on the participants, such as their age, gender, geographic origins, and the specific communities in which they resided. This approach aimed to comprehend the environment that allowed the riots to occur, particularly given that the violence was fueled by a false premise circulating on the internet that the attacker was a Muslim immigrant. According to Duncan, the initial rumor originated from a small social media aggregator website with only a few thousand followers. Despite its limited audience, the claim became highly viral, demonstrating the rapid speed at which misinformation can propagate within today’s information ecosystem. The research team analyzed approximately 51,000 posts, a relatively small sample size for social media studies. Their findings indicated that these groups were moderated primarily by individuals of middle age or retirement age, with a significant number being women. Duncan noted that many of these moderators genuinely believed the information they shared. She argued that for the majority of participants, this did not constitute intentional disinformation or manipulation in the traditional sense. Instead, these individuals perceived the claims as factual and repeated them uncritically because they lived in an informational environment that mirrored their existing beliefs. Duncan cautioned against attributing all responsibility solely to social media platforms. She described the situation as part of a broader societal issue where social media serves merely as a tool exploited by malicious actors. She observed that the current era of misinformation did not emerge in a vacuum. The anti-establishment sentiment identified in their research appeared deeply rooted in wider social problems, driven by legitimate concerns among people regarding their future and the future of their children. Based on the study, Duncan concluded that the riots resulted from years of regulatory neglect concerning social media platforms. These platforms have accumulated increasing power within societies and democracies without adequate oversight. She suggested that the views expressed by these users on social media likely mirror conversations held in private settings, such as around kitchen tables. Consequently, politicians often adopt this rhetoric due to the inherently short-sighted nature of politics. In a separate incident highlighting the dynamics of online interaction, a young man in Osijek, Croatia, posted in the Facebook group Spotted: Osijek-Slavonija after his first date fell apart over a transportation dispute. The man, who knew the woman from university, had planned to pay for dinner and drinks in the city’s fortress area. When the woman mentioned she might take a taxi because it was cold and windy, he objected, arguing she should save money by taking the tram. He checked a public transit tracking website, found a tram arriving in three minutes, and sent her a screenshot of the map with the tram circled in red. His message included a comment telling her not to act like a lady and to save some money since there would be no waiting time. The woman responded briefly, thanking him for the information, before ceasing communication. She did not appear for the date. After waiting for half an hour, the man called her. She explained that his focus on her transport costs before they had even met, combined with his unsolicited financial advice via screenshots, made the relationship untenable. She blocked him afterward. The man admitted he was confused by her reaction, believing he had done her a favor by saving her five euros and ensuring she did not wait in the winter weather. He asked the group whether he was wrong for valuing money or whether modern expectations require accepting others’ spending decisions without objection. Most comments in the thread sided with the woman. One user labeled him a control freak, while another noted that she correctly identified behavior she disliked early on. Commenters pointed out that she was not spending his money and had not asked him to pick her up, criticizing his attempt to dictate how she managed her own funds.