Google has announced a new tool for keeping track of news involving hate crimes in America called the Documenting Hate News Index. Built in collaboration with ProPublica, who originally got together with Google on the Documenting Hate project, the specialized news tracker is made to find and archive news stories about hate crimes, then pull key data from them in order to build a dataset about hate crimes in America. The new tool accomplishes this indexing through machine learning, the Alphabet-owned company revealed. The AI behind it all is tuned to comb through any articles containing key terms like “hate,” “terrorism,” and “hate crime” and use Google’s own natural language processing protocols, the same ones that power Google Assistant and other speech tools, to figure out whether the story it’s looking at truly pertains to a hate crime, and that the crime in question is not already somewhere in the database.
The tool was created in response to the difficulty that journalists can have chronicling hate crimes using online news resources. Even if not for that fact, however, gathering cohesive, holistic, and comprehensive data sets regarding hate crime reports in America would take unity between all news staff reporting on them, and a lot of coordination. An AI, on the other hand, can do the same work much faster, and total up meaningful data extracted from stories to add to a self-maintained database, taking out the potential for individual and collaborative human error.
There has previously been talk of whether AI could replace certain fields that require insight, decision-making, and emotion that only humans can currently muster. While the field of journalism may seem ripe for an AI takeover due to the stark and cold nature of the task of gathering, compiling, and interpreting data, AI solutions are not yet at the stage where they could conceivably provide relevant background, side content, and insight on a given story, depending on the type of data at hand. In this case, while the Documenting Hate News Index can gather up data on hate crimes from existing news stories, it cannot craft its own by gleaning details from police reports, citizens’ social media pages, and the like. A tool like this may not be a journalistic revolution, but it does promise to provide detailed and cumulative insight into hate crimes in America.