Auditing Elon Musk's Impact on Hate Speech and Bots

Daniel Hickey,Matheus Schmitz,Daniel Fessler,Paul Smaldino,Goran Muric,Keith Burghardt

DOI: https://doi.org/10.1609/icwsm.v17i1.22222

2024-01-29

Abstract:On October 27th, 2022, Elon Musk purchased Twitter, becoming its new CEO and firing many top executives in the process. Musk listed fewer restrictions on content moderation and removal of spam bots among his goals for the platform. Given findings of prior research on moderation and hate speech in online communities, the promise of less strict content moderation poses the concern that hate will rise on Twitter. We examine the levels of hate speech and prevalence of bots before and after Musk's acquisition of the platform. We find that hate speech rose dramatically upon Musk purchasing Twitter and the prevalence of most types of bots increased, while the prevalence of astroturf bots decreased.

Social and Information Networks

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to evaluate the changes in hate speech and bots on the platform after Elon Musk's acquisition of Twitter. Specifically, the researchers focused on the following aspects:

1. **Changes in hate speech**: The researchers wanted to know whether there were significant changes in the level of hate speech on the platform before and after Musk's acquisition of Twitter. They measured this change by analyzing tweet data within a specific time period, especially tweets containing hate - related keywords.

2. **Changes in bot accounts**: In addition to hate speech, the researchers also focused on whether the activity level of bot accounts (bots) on the platform had changed. They used the Botometer tool to evaluate the distribution of different types of bot accounts before and after the acquisition.

### Research background

Musk acquired Twitter on October 27, 2022, and became its new CEO. He proposed a series of reform measures, including reducing content moderation and relaxing restrictions on spam bots. These measures have raised concerns about platform security and content health, especially regarding the increase in hate speech.

### Research methods

1. **Extraction of hate tweets**:
- Use a vocabulary related to hate groups to identify a series of hate - related keywords.
- Collect historical tweets containing these keywords from the Twitter API.
- Use Google's Perspective API to further filter out toxic tweets.

2. **Activity trails of hate - using users**:
- Select users who posted hate tweets between October 1, 2022 and November 29, 2022, and collect all of their tweets during this period.
- Calculate the proportion of hate tweets posted by each user every day to track the trend.

3. **Bot detection**:
- Use the Botometer tool to evaluate the bot scores of accounts that post hate tweets and other control accounts.
- Compare the distribution of different types of bot accounts before and after the acquisition.

### Main findings

1. **Increase in hate speech**:
- After Musk's acquisition of Twitter, the level of hate speech increased significantly, especially with a sharp increase in the early days of the acquisition.
- Overall, the number of hate tweets increased significantly in the second half of 2022.

2. **Changes in bot accounts**:
- The number of most types of bot accounts increased, especially spam bots and fake followers.
- However, the number of "astroturf bots" decreased.

### Conclusion

The research results show that after Musk's acquisition of Twitter, both hate speech and bot activities on the platform have increased. Although the specific reasons for these changes are still unclear, the researchers speculate that this may be related to Musk's reduction in content moderation and the large number of platform employees leaving. These findings pose challenges to the future healthy development of Twitter and require further research and policy adjustments to address these problems.