Figure 9 - uploaded by Emil O. W. Kirkegaard
Content may be subject to copyright.
Journalists' relative support for parties by political ideology. Red diamonds correspond to the weighted median for each tag in the log10 RR metric.
Source publication
We gathered survey data on journalists' political views in 17 Western countries. We then matched these data to outcomes from national elections, and constructed metrics of journalists' relative preference for different political parties. Compared to the general population of voters, journalists prefer parties that have more left-wing positions over...
Context in source publication
Citations
... Other sources of data likely used for LLMs' training are news media articles and academic papers. Research has shown that in the U.S., the U.K., and many other Western nations, there are more left-leaning than right-leaning journalists [19], [20], [21]. Similarly, academics also tend to lean, on average, left-of-center [22], [23], [24]. ...
Political biases in Large Language Model (LLM)-based artificial intelligence (AI) systems, such as OpenAI's ChatGPT or Google's Gemini, have been previously reported. While several prior studies have attempted to quantify these biases using political orientation tests, such approaches are limited by potential tests' calibration biases and constrained response formats that do not reflect real-world human-AI interactions. This study employs a multi-method approach to assess political bias in leading AI systems, integrating four complementary methodologies: (1) linguistic comparison of AI-generated text with the language used by Republican and Democratic U.S. Congress members, (2) analysis of political viewpoints embedded in AI-generated policy recommendations, (3) sentiment analysis of AI-generated text toward politically affiliated public figures, and (4) standardized political orientation testing. Results indicate a consistent left-leaning bias across most contemporary AI systems, with arguably varying degrees of intensity. However, this bias is not an inherent feature of LLMs; prior research demonstrates that fine-tuning with politically skewed data can realign these models across the ideological spectrum. The presence of systematic political bias in AI systems poses risks, including reduced viewpoint diversity, increased societal polarization, and the potential for public mistrust in AI technologies. To mitigate these risks, AI systems should be designed to prioritize factual accuracy while maintaining neutrality on most lawful normative issues. Furthermore, independent monitoring platforms are necessary to ensure transparency, accountability, and responsible AI development.
... The theoretical background of this approach is that any knowledge worker will consciously or unconsciously inject their own political values into their work. Prior work on journalists found a marked preference for left-wing and green (environmental) politics and aversion to right-wing populism (Kirkegaard et al., 2021). Because Wikipedia editors essentially function as journalists, translating primary and secondary sources into summaries for the general audience, the political leanings of Wikipedia editors are expected to be similar to those of journalists. ...
... Overall we find a left-wing skew, with 81 % being left-wing, 15 % right-wing, and 5 % centrist. These findings are similar to results for journalists (Kirkegaard et al., 2021). We interpret the results in line with a left-wing dominated system of information flow in society. ...
We scraped user pages of 7,739 Wikipedia editors. Of these, 224 users positioned themselves politically using the semi-standardized "userboxes". Based on this sample, Wikipedia editors' views had a strong tilt towards the left. The results are congruent with the political leanings of related occupations, such as journalists and academics.
... The lack of ideological diversity within the journalism profession has been described before in the United States [28], the U.K [29]. and 15 others Western countries [30]. That is, most journalists' political orientation sits left-of-center in comparison to the general population. ...
Previous research has identified a post-2010 sharp increase of terms used to denounce prejudice (i.e. racism, sexism, homophobia, Islamophobia, anti-Semitism, etc.) in U.S. and U.K. news media content. Here, we extend previous analysis to an international sample of news media organizations. Thus, we quantify the prevalence of prejudice-denouncing terms and social justice associated terminology (diversity, inclusion, equality, etc.) in over 98 million news and opinion articles across 124 popular news media outlets from 36 countries representing 6 different world regions: English-speaking West, continental Europe, Latin America, sub-Saharan Africa, Persian Gulf region and Asia. We find that the post-2010 increasing prominence in news media of the studied terminology is not circumscribed to the U.S. and the U.K. but rather appears to be a mostly global phenomenon starting in the first half of the 2010s decade in pioneering countries yet largely prevalent around the globe post-2015. However, different world regions’ news media emphasize distinct types of prejudice with varying degrees of intensity. We find no evidence of U.S. news media having been first in the world in increasing the frequency of prejudice coverage in their content. The large degree of temporal synchronicity with which the studied set of terms increased in news media across a vast majority of countries raises important questions about the root causes driving this phenomenon.
... Another hypothesis to explain the media trends described in this work could be the increasing intellectual homogeneity in newsrooms. The lack of political diversity in the journalism profession has been described before in the United States (Weaver et al., 2019), the UK (Journalists in the UK, n.d.) and 15 others Western countries (Kirkegaard et al., 2021). That is, most journalists' political orientation sits left of center in comparison to the general population. ...
Previous research has identified a post-2010 sharp increase of words used to denounce prejudice (i.e. racism, sexism, homophobia, Islamophobia, anti-Semitism, etc) in US and UK news media content. Some have referred to these institutional trends and related shifts in US public opinion about increasing perceptions of prejudice severity in society as the Great Awokening. Here, we extend previous analysis to the global media environment. Thus, we quantify the prevalence of prejudice-denouncing terms and social justice associated terminology (diversity, inclusion, equality, etc) in over 98 million news and opinion articles across 124 popular news media outlets from 36 countries representing 6 different world regions: English-speaking West, continental Europe, Latin America, sub-Saharan Africa, Persian Gulf region and Asia. We find that increasing prominence in news media of so-called wokeness terminology is a global phenomenon starting early post-2010 in pioneering countries yet mostly worldwide ubiquitous post-2015. Still, different world regions emphasize distinct types of prejudice with varying degrees of intensity. Surprisingly, the United States news media does not appear to have been the pioneer in embedding prejudice and social justice loaded terminology in their content. We also note that state-controlled news media from Russia, China and Iran might be leveraging wokeness terminology as a geopolitical propaganda weapon to mock, destabilize or criticize Western adversaries. The large degree of temporal synchronicity with which wokeness terminology emerged in news media worldwide raises important questions about the root causes driving this phenomenon.
