With contributions by Steven Finkel, Daniel Wallace Professor of Political Science, University of Pittsburgh
Firat Duruşan, PhD, Center for Computational Social Sciences, Koç University
Erdem Yörük, Director, Center for Computational Social Sciences, Koç University
Isik Topçu, MA, Center for Computational Social Sciences, Koç University
Melih Can Yardi, MA, Center for Computational Social Sciences, Koç University
Amanda Zaner, MID, GSPIA, University of Pittsburgh
Introduction
Women, and women of color in particular, face numerous challenges when running for political office in the U.S. These include attacks they are subject to in various online spaces that, like their peers, they must use to campaign and promote their work. Theseattacksoften aim to undermine and prevent women’s participation in politics. Previous research by CDT found that women of color Congressional candidates in the 2020 U.S. election were more likely to be subjected to violent and sexist abuse, and mis- and disinformation on X/Twitter compared to other candidates. These forms of abuse might contribute to the underrepresentation of women of color in politics, and may also undermine the effectiveness of the US democratic system in reflecting the interest and priorities of all voters in policy-making.
In this research brief, we turn to the 2024 U.S. elections to examine the nature of offensive speech and hate speech that candidates running for Congress are subject to on the social media platform X (formerly Twitter), which remains an important forum for political candidates. More specifically, we compare the levels of offensive speech and hate speech that different groups of Congressional candidates are targeted with based on race and gender, with a particular emphasis on women of color. We also examine these factors for U.S. Vice President Kamala Harris as a woman of color and presidential candidate.
For purposes of this research, we identified all tweets posted between May 20 and August 23, 2024 that mentioned any candidate running for Congress with one or more accounts on X (a total of 1031 candidates). This resulted in a dataset of over 800,000 tweets. Additionally, we examined tweets that mentioned Harris during this period. Using several fine-tuned language models, we identified tweets that contained offensive speech or hate speech (see the methods section in the report for more details) about the candidate. We define offensive speech as words or phrases that demean, threaten, insult, or ridicule a candidate. We define hate speech as a subset of offensive speech where specific reference is made to someone’s identity including race, gender, sexual orientation, or religion. Our findings show that women of color and African American women candidates in particular are subject to more offensive speech overall, and specifically to more hate speech, than other candidates.
[CONTENT WARNING – Some of the examples in this report include profanity and threats that some may find offensive or triggering.]
On September 15, 2026, the Center for Democracy & Technology and Electronic Frontier Foundation filed an amicus brief along with the Wikimedia Foundation in Meta Platforms v. Bonta, challenging California’s SB 976.
Not All Guardrails Are Created Equal: Comparing Content Safety and Copyright Filtering
As courts and policymakers work through questions about chatbot liability, they should be wary of analogies that flatten meaningful technical differences. Copyright filtering and safety intervention share real challenges around ambiguity and evasion, but they diverge in what each control must assess, how each manifests over the course of a conversation, and how much can be verified from the outside.
CDT Joins Letter Defending First Amendment Rights Amid Attack on Center for American Progress
The Center for Democracy & Technology joined more than 150 civil society, advocacy, and nonpartisan organizations in a joint statement led by Nonprofits Together supporting the Center for American Progress amid a recent legal action by the Trump administration. The statement emphasizes that independent organizations have a First Amendment right to publish evidence-based analysis and critique government policies, regardless of which administration is in power.
Op-Ed: France’s Constitutional Council Ruling Tests the Limits of Social Media Bans
Read our analysis the legal and policy significance of the Constitutional Council’s decision on ongoing social media age restriction and age assurance debates ahead of the most-anticipated EU-wide approach on child online safety due by December 2026.
