Read the passage and mark the letter A, B, C or D on your answer sheet to indicate the best answer to each of the following questions from 3...
Đề bài
Read the passage and mark the letter A, B, C or D on your answer sheet to indicate the best answer to each of the following questions from 33 to 40.
The deeply polarizing nature of social media is not a consequence of deficient algorithms but an inevitable outcome of the platforms' fundamental architecture, a new study utilizing AI-generated users has concluded. The research indicates this malady cannot be solved by superficial adjustments and will persist unless the core mechanics of online communication are fundamentally reimagined. The issue appears inextricable from the very concept of posting, reposting, and following, which inherently foments problematic social dynamics.
To interrogate this hypothesis, researchers at the University of Amsterdam, led by Petter Törnberg, devised a simulation with 500 AI chatbots. These agents were programmed to emulate a spectrum of US political beliefs based on established survey data. They were then instructed to interact on a bare-bones social network, a platform deliberately divested of all algorithms, advertisements, and engagement-maximization features. The explicit goal was to observe what emergent behaviors would manifest in such a neutral environment, absent the typical corporate incentives.
The outcomes were immediate and stark. The AI agents instinctively gravitated toward and followed users with whom they shared political affiliations. Bots expressing more partisan or extreme views consistently amassed more followers and reposts, effectively monopolizing the network's attention. The team then tested six potential interventions to counter these behaviors, including a chronological feed and amplifying opposing views. None of the remedies proved effective; most yielded negligible changes, while some, like hiding biographies, actually exacerbated polarization.
Researchers confronted a frustrating trade-off: alterations that reduced user inequality inadvertently made extreme posts more popular. Törnberg, who had presumed algorithms were the primary culprit, admitted the findings were a surprise. He concluded that “wiggling with algorithms” is woefully insufficient. The simulation strongly suggests that a “fundamental rethinking” of the very structure of online interaction is imperative to prevent these digital spaces from distorting public and political discourse.
Question 33: The word foments in paragraph 1 is closest in meaning to ______.
A. reveals B. incites C. creates D. reflects