Analyzing Accessibility of Wikipedia Projects Around the World

Clark, Justin; Faris, Robert; Jones, Rebekah Heacock

doi:10.2139/ssrn.2951312

Cited by 6 publications

(10 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The adoption of HTTPS by highly visited websites such as Wikipedia has resulted in greater accessibility for those sites in many cases. 40 As of August 2011, Saudi Arabia was blocking specific Wikipedia entries such as one on the theory of evolution and individual Twitter accounts such as those of Egyptian activist Wael Ghonim and Gamal Eid, the director of a Cairo-based regional human rights NGO who often posts tweets critical of the record of freedom of expression in Saudi Arabia and other Arab countries. The website of Eid's human rights organization remains blocked.…”

Section: Internet Monitormentioning

confidence: 99%

“…Furthermore, the 2017 study found that individual articles subject to censorship in the past have been receiving increased levels of traffic since Wikipedia transitioned to HTTPS in June 2015, suggesting that the transition has made it more difficult for Iran to censor selected Wikipedia content. 44 Other government censors choose to block entire domains and platforms. For example, online activists, writers, and blocked websites in Egypt resorted to using the online publishing platform medium.com to disseminate their content, exploiting the fact that the censors cannot ban individual accounts since Medium uses HTTPS by default.…”

Section: Internet Monitormentioning

confidence: 99%

See 1 more Smart Citation

The Shifting Landscape of Global Internet Censorship

et al. 2017

Self Cite

View full text Add to dashboard Cite

Section: Internet Monitormentioning

confidence: 99%

Section: Internet Monitormentioning

confidence: 99%

The Shifting Landscape of Global Internet Censorship

et al. 2017

Self Cite

View full text Add to dashboard Cite

“…The test set contains 5,291 headlines with positive sentiment, 3,913 with negative sentiment, and 3,424 with neutral sentiment. 10 We preprocess the news headlines by removing punctuation, numbers, special characters, the names of the news agency (if they appear on the headline), and duplicated headlines. To convert the news headlines into input for machine learning models, we first use a Chinese word segmentation tool to segment each news headline into a sequence of words.…”

Section: Methodsmentioning

confidence: 99%

“…9 16 duplicated news headlines are dropped, resulting in 4,984 headlines in total. 10 41 duplicated news headlines are dropped, resulting in 12,628 headlines in total. 11 Because headlines with neutral labels are more noisy and given the difficulty of training a three-class classifier with limited training data, we report results in the main text based on models that are trained with only positive and negative headlines.…”

Section: Methodsmentioning

confidence: 99%

“…Government censorship of social media, news, and websites directly affects large corpuses of text by blocking users' access, deleting individual messages, adding content through propaganda, or inducing self-censorship through intimidation and laws [11,19,20,22,25,32,34]. While Wikipedia's global reach makes it an attractive corpus for training models in many different languages, Wikipedia has also been periodically censored by many governments, including Iran, China, Uzbekistan, and Turkey [10]. China has had the most extensive and long-lasting censorship of Wikipedia.…”

Section: Censorship Of Chinese Language Wikipedia and Implications Fo...mentioning

confidence: 99%

See 1 more Smart Citation

Censorship of Online Encyclopedias

Yang

Roberts

2021

Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency

View full text Add to dashboard Cite

While artificial intelligence provides the backbone for many tools people use around the world, recent work has brought to attention that the algorithms powering AI are not free of politics, stereotypes, and bias. While most work in this area has focused on the ways in which AI can exacerbate existing inequalities and discrimination, very little work has studied how governments actively shape training data. We describe how censorship has affected the development of Wikipedia corpuses, text data which are regularly used for pre-trained inputs into NLP algorithms. We show that word embeddings trained on Baidu Baike, an online Chinese encyclopedia, have very different associations between adjectives and a range of concepts about democracy, freedom, collective action, equality, and people and historical events in China than its regularly blocked but uncensored counterpart -Chinese language Wikipedia. We examine the implications of these discrepancies by studying their use in downstream AI applications. Our paper shows how government repression, censorship, and self-censorship may impact training data and the applications that draw from them. CCS CONCEPTS• Computing methodologies → Supervised learning by classification; • Information systems → Content analysis and feature selection; • Social and professional topics → Political speech.

show abstract

Is remote measurement a better assessment of internet censorship than expert analysis? Analyzing tradeoffs for international donors and advocacy organizations of current data and methodologies

Fletcher

Hayes-Birchler²

2023

Data & Policy

View full text Add to dashboard Cite

Donor organizations and multilaterals require ways to measure progress toward the goals of creating an open internet, and condition assistance on recipient governments maintaining access to information online. Because the internet is increasingly becoming a leading tool for exchanging information, authoritarian governments around the world often seek methods to restrict citizens’ access. Two of the most common methods for restricting the internet are shutting down internet access entirely and filtering specific content. We conduct a systematic literature review of articles on the measurement of internet censorship and find that little work has been done comparing the tradeoffs of using different methods to measure censorship on a global scale. We compare the tradeoffs between measuring these phenomena using expert analysis (as measured by Freedom House and V-Dem) and remote measurement with manual oversight (as measured by Access Now and the OpenNet Initiative [ONI]) for donor organizations that want to incentivize and measure good internet governance. We find that remote measurement with manual oversight is less likely to include false positives, and therefore may be more preferable for donor organizations that value verifiability. We also find that expert analysis is less likely to include false negatives, particularly for very repressive regimes in the Middle East and Central Asia and therefore these data may be preferable for advocacy organizations that want to ensure very repressive regimes are not able to avoid accountability, or organizations working primarily in these areas.

show abstract

Analyzing Accessibility of Wikipedia Projects Around the World

Cited by 6 publications

References 7 publications

The Shifting Landscape of Global Internet Censorship

The Shifting Landscape of Global Internet Censorship

Censorship of Online Encyclopedias

Is remote measurement a better assessment of internet censorship than expert analysis? Analyzing tradeoffs for international donors and advocacy organizations of current data and methodologies

Contact Info

Product

Resources

About