And this is what science says
Posted: Sun Jan 19, 2025 6:12 am
… Ranking algorithms are indeed biased and "flawed" By now, there have been many studies of ranking methods implemented by network platforms. Let's consider one of them - the work "Network Ranking in Social Systems: Three Problems" published by employees of the University of Zurich (Switzerland), the Institute of Basic and Frontier Sciences (part of the University of Electronic Science and Technology of China) and the Alibaba Research Center at Hangzhou Normal University (yes, it is related to that very colossal Chinese company).
Of the three problems described in the study, two are directly relevant list of cyprus whatsapp phone numbers to the conversation: “Ranking algorithms influence the behavior of agents ( users and content creators – author’s note ), leading to potentially perverse feedback mechanisms and unhealthy systemic consequences.” The authors of the study point out that people have previously hoped that ranking and forecasting algorithms would be able to provide an objective assessment of an agent’s value (whether it’s the quality of a cultural product, a person’s talent, or the relevance of a web page).
They say that people’s expert opinions cannot be trusted because they are subjective and dependent on biases and social factors, while trusting the opinions of algorithms is potentially more reasonable behavior. However, in practice it turned out that blindly accepting the results of the algorithms is misleading. Quote: “…we argue that an appropriate dose of caution is needed here in interpreting the results… as a signal of quality or talent.
Of the three problems described in the study, two are directly relevant list of cyprus whatsapp phone numbers to the conversation: “Ranking algorithms influence the behavior of agents ( users and content creators – author’s note ), leading to potentially perverse feedback mechanisms and unhealthy systemic consequences.” The authors of the study point out that people have previously hoped that ranking and forecasting algorithms would be able to provide an objective assessment of an agent’s value (whether it’s the quality of a cultural product, a person’s talent, or the relevance of a web page).
They say that people’s expert opinions cannot be trusted because they are subjective and dependent on biases and social factors, while trusting the opinions of algorithms is potentially more reasonable behavior. However, in practice it turned out that blindly accepting the results of the algorithms is misleading. Quote: “…we argue that an appropriate dose of caution is needed here in interpreting the results… as a signal of quality or talent.