An open applied research project on detecting political or other types of social media influence via bots.
Switch branches/tags
Nothing to show
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.
Failed to load latest commit information.

Title: Social media botnet detection and analysis Tagline: Analyze political botnet activity on Twitter and develop effective counter-measures. Date: October 2016 Category: Applied Research Mailing list: Contact:

Problem description

Develop a model capable of reliably flagging politically-motivated botnets on social media (e.g. Twitter) that are engaging in propaganda and disinformation. Use this model to investigate the scale and effectiveness of botnet-mediated opinion manipulation campaigns. Devise effective counter-measures.

Why this problem matters

A major, near-term threat posed by artificial intelligence is the deployment of effective disinformation and opinion manipulation campaigns on social media. They may come from nation-states (whether targeted at foreign powers or at their own citizens) and may be extremely well-funded, well-staffed, and technologically advanced. There is already considerable evidence that Russia is using low-tech versions of such tactics to attempt to manipulate the American public. AI will soon allow to scale and refine these techniques into a powerful weapon. The only realistic counter-measure would be the ability to reliably detect and flag non-human activity and opinion manipulation attempts on social media.


Collect tweets via the Twitter API.