Facebook’s Impact on Politics Questioned in New Research
New research suggests that Facebook’s algorithms may not have as significant an impact on users’ political beliefs and behavior as previously thought. Four research papers released this week found little evidence that the information amplified by Facebook’s algorithms significantly influenced users’ political opinions. The findings challenge claims made by Meta, Facebook’s parent company, that its platforms do not contribute to political divisiveness.
Meta’s President of Global Affairs, Nick Clegg, stated that the experimental findings add to a growing body of research showing that Meta’s platforms alone do not cause harmful political polarization or have substantial effects on political outcomes. However, the researchers involved in the studies disagree with Clegg’s conclusions. Talia Stroud, a communications professor at the University of Austin and part of a team investigating anonymized Facebook data, stated that the researchers’ conclusions do not align with Clegg’s statements.
The papers, published in Nature and Science, indicate that political news shared and posted on Facebook does not significantly sway users’ political beliefs. While these findings challenge the notion that Facebook has a substantial impact on political polarization, they do not definitively settle the debate about social media’s influence on democracy and political attitudes.
A total of 16 papers will be released on this subject, aiming to advance society’s understanding of the complex relationship between social media and politics. Meta’s CEO, Mark Zuckerberg, discussed the company’s licensing model during their recent earnings call. The license for Meta’s Llama 2 large language model, billed as open source, imposes restrictions on major cloud providers who generate significant revenue by reselling Meta’s services. The move has been criticized by developers, arguing that the license lacks true openness.
While Meta views the licensing model as a way to earn revenue from large cloud players, critics argue that it limits the model’s accessibility to developers. Revising the licensing restrictions could help Meta attract more developers and compete with other models like those developed by OpenAI and Google.
In other news, Google DeepMind announced the introduction of Robotic Transformer 2 (RT-2). This system combines computer vision, natural language processing, and robotics to enhance the agility and adaptability of robots. RT-2 enables robots to perform tasks beyond a predefined set and apply their skills to new objects or interpret unfamiliar commands. The vision-language model allows the robotic gripper to recognize and manipulate previously unseen objects, making rudimentary reasoning possible.
DeepMind’s innovative approach to robotics showcases enhanced generalization capabilities, semantic understanding, and visual interpretation. By predicting the next action instead of the next word, RT-2 exhibits flexibility and expands the potential applications of robots.
These developments in the fields of artificial intelligence and social media raise important questions about their implications for society, such as the influence of online platforms on political discourse and the future of robotics and automation. The ongoing research and advancements in these areas will continue to shape our understanding and usage of technology.
