What percentage of Twitter accounts are fake?

1. Summary of the results

The percentage of fake Twitter accounts is a topic of ongoing debate, with varying estimates provided by different sources ^{[1] [2] [3]}. According to Twitter's filing, less than 5% of its monetizable daily active users are false or spam accounts ^[1]. However, a report from Cyabra found that 15% of accounts praising former President Donald Trump and criticizing President Joe Biden are fake, and 7% of accounts praising President Biden and criticizing Trump are also fake ^[2]. Another source estimates that between a third and two thirds of Twitter accounts are bots, but does not provide a specific percentage ^[3]. Twitter itself reports removing over 1 million spam accounts each day, with spam accounts being well under 5% of users who are served advertising ^[4].

2. Missing context/alternative viewpoints

Some sources do not provide a direct estimate of the percentage of fake Twitter accounts, but highlight the issue of supersharers who spread fake news, which could be related to the presence of fake accounts ^[5]. The challenges of detecting social media bots and the methods used to identify them are also discussed ^[6]. Additionally, a Reddit discussion shares users' experiences with spam bots on Twitter, with some users estimating 'near 100% bot saturation' in certain trending hashtags ^[7]. The launch of an initiative to eliminate spam and fake accounts by Twitter, now referred to as X, is also reported ^[8]. These alternative viewpoints suggest that the issue of fake accounts on Twitter is complex and multifaceted, and that different methods of detection and estimation may yield varying results ^{[1] [2] [3]}.

• The presence of fake accounts can have significant implications for the spread of misinformation and the integrity of online discourse ^{[5] [6]}.
• The use of large language models to detect social media bots can be both helpful and harmful ^[3].
• The effectiveness of Twitter's efforts to remove spam accounts is a topic of ongoing discussion ^{[4] [7]}.

3. Potential misinformation/bias in the original statement

The original statement asks for a specific percentage of fake Twitter accounts, but the provided analyses suggest that this information is not readily available or is subject to varying estimates ^{[1] [2] [3]}. Some sources may benefit from exaggerating or downplaying the prevalence of fake accounts, such as Twitter itself, which may have an interest in presenting a lower estimate ^{[1] [4]}. On the other hand, sources that highlight the issue of supersharers and fake news may benefit from emphasizing the severity of the problem ^{[5] [6]}. Additionally, the methodology used to detect and estimate fake accounts can also be a source of bias, as different methods may yield varying results ^{[1] [2] [3]}. Overall, it is essential to consider multiple sources and viewpoints when assessing the percentage of fake Twitter accounts ^{[1] [5] [2] [3] [6] [8] [4] [7]}.