Humane Intelligence Algorithmic Bias Bounty 2
CHALLENGE 2
Advancing AI to uncover hidden extremist propaganda
Humane Intelligence is thrilled to be launching the second of 10 "algorithmic bias bounty" programs, which will be unfolding over the coming year. With the support of Google.org, we are building themed programs that aim to build community and professionalize the practice of algorithmic assessment. This second challenge is in partnership with Revontulet, a human-centric company that provides intelligence and analysis to help clients around the globe mitigate risk and prevent harm caused by terrorism and violent extremism.
The Challenge
Our second Challenge in partnership with Revontulet will focus on counterterrorism in computer vision (CV) applications. Centering on far-right extremist groups in Europe, the Nordic region, and beyond and the rise of AI-generated content, the goal is to train a CV model to understand the ways in which hateful image-propaganda can be disguised and manipulated to evade detection on social media platforms. Participants will be tasked with developing an image identification solution capable of extracting, discerning, and interpreting context related to counterterrorism. The outcome of this challenge will assess the robustness and reveal potential gaps in the feasibility and efficacy of applying machine learning as a tool in counterterrorism efforts. Challenge participants compete for $10,000 in prizes - $4000 for intermediate and $6000 for advanced challenge winners. In addition, your work may be featured in Revontulet's suite of counterterrorism solutions.
Important considerations:
-
Due to the sensitive nature of this topic, we are only hosting intermediate and advanced challenges to skew towards more experience.
-
Participants will need to sign this waiver in order to receive the dataset. In this waiver you attest to use the data appropriately for the challenge.
-
We are only accepting applicants who are age 18+, given the sensitivity of the subject.
-
If you experience distress during the course of this challenge, please stop working on the challenge and email us at hi@humane-intelligence.org, or message us on our Discord channel. We will provide mental health resources.
Intermediate
The task is to build an unsupervised machine learning model that groups unlabeled images into 2 clusters to identify whether an image contains extremist content or not.
Advanced
Building on top of the intermediate challenge, the task is to create adversarial examples to test the robustness of your unsupervised model.
The Dates
September 26, 2024
Challenge opens.
October 28, 2024
First day of accepting submissions to the competition.
November 7, 2024
Competition closes.
November 29, 2024
Winners announced.
Sign up to participate!
You are welcome to get involved.
Sign up here to get on our mailing list and to hear from us about competition-specific logistics
