Facebook has launched a competition to help the social media giant erase "Hateful Memes" from its site — and is offering $100,000 in prizes for the programs that can accomplish the task.
What are the details?
Of course, it's easy for programs to spot actual hate speech and delete it. The Facebook AI team defines "hate speech" as:
"A direct or indirect attack on people based on characteristics, including ethnicity, race, nationality, immigration status, religion, caste, sex, gender identity, sexual orientation, and disability or disease. We define attack as violent or dehumanizing (comparing people to non-human things, e.g. animals) speech, statements of inferiority, and calls for exclusion or segregation. Mocking hate crime is also considered hate speech."
"This definition mirrors the community standards on hate speech employed by Facebook, and is intended to provide an actionable classification label: if something is hate speech according to this definition, it should be taken down; if not, even if it is distasteful or objectionable, it is allowed to stay," the competition page states.
But then there's the problem of hateful memes being created by combining otherwise harmless phrases and otherwise harmless images:
Image source: DrivenData video screenshot
Which is why the tech giant is instructing participants in its challenge to create artificial intelligence that can detect text and images used together as a hateful meme.
"We continue to make progress in improving our AI systems to detect hate speech and other harmful content on our platforms, and we believe the Hateful Memes project will enable Facebook and others to do more to keep people safe," the competition page also said.