Bots are increasingly interfering in political elections and economic debates. Italian scientist Stefano Cresci analyzed 236 scientific papers devoted to bots. He outlined the patterns he found in an article. Evgeny Kolesnikov translated the material and invites all Likeni readers to learn more about social bots and how to detect them.
On the morning of November 9, 2016, the world woke up to the shocking result of the US presidential election: Donald Trump had become the 45th President of the United States of America. An unexpected event that continues to have repercussions around the world.
Today, we know that social bots – automated social media accounts that imitate humans – played a central role in spreading discord and misinformation, possibly contributing to Trump’s victory.[16][19]
Key points
- Social bots have been studied for a long time, but they email data still remain an unsolved problem in the online ecosystem. Over time, several trends have emerged in their detection. The latest and most promising advancement in this area is related to group-based detectors.
- Deception detection is inherently adversarial. Applying adversarial machine learning can give us an edge in combating all forms of online manipulation and automation.
- Recent advances in AI and computing (e.g. deepfakes) make individual bots indistinguishable from users. Future efforts should focus on measuring the degree of inauthentic coordination rather than attempting to classify the nature of individual accounts.
Following the 2016 US election, the world began to wake up to the seriousness of social media deception. Following the Trump exploit, we witnessed a stark dissonance between the marketo integrations: why do you need them? many efforts being made to detect and remove bots and the growing influence of these malicious actors on society.[27][29] This paradox raises the question: what strategies should we employ to stop the social bot pandemic?
In the lead-up to the 2020 US elections, this question seems more important than ever, particularly in light of reports of thousands of AI-powered accounts rigging election debates [a].
What has struck social, political, and economic analysts since 2016—deception and automation—has been a subject of scholarly study since at least 2010. In this paper, we briefly european union phone number review the first decade of research in social bot detection. Using a longitudinal analysis, we discuss the main trends in counter-bot research and the factors that make this never-ending battle so difficult. Using lessons learned from our analysis, we propose potential innovations that could give us an edge against deception and manipulation.