The difference between May 13 and May 15 is the occurrence of a single data point, the name Payton S. Gendron. With that name in hand, police and reporters had no trouble quickly gathering records on racist threats and racist theories, fascinated by mass shootings. Prior mass shootings, police intervention, target-specific planning and gun purchases all point directly to the deadly mass-casualty event that took place on May 14.
Tops Supermarket in Buffalo was the scene of a terrorist attack, a massacre planned by a politically motivated psychopath. Another such attack occurred on the same weekend in Orange County, California. A man with political grievances attacked, with muskets and guns, a church frequented by Taiwanese parishioners, only one person was killed thanks to the quick intervention of bystanders.
Gun control could be the answer to all types of gun crime, from armed robbery and domestic murder to downtown Chicago gang shootings and style massacres. Events are planned like in Orange and Buffalo counties, if gun control goes far enough — that is, if it takes away the serious American right to buy and own a gun.
Even Democrats, the pro-gun control party, could not find support in their own caucus for restrictions that could have a significant effect on such crimes. By contrast, Americans have made it very clear, with their tolerance for everything from online tracking to E-ZPass, from traffic cameras and license plate readers to in-store facial recognition, that they willing to tolerate a lot of intrusion if passive surveillance.
A gun control revolution is not going to happen. Even those who support such a revolution today derive primarily from a need to show desperation and disdain for fellow Americans who place a higher value on their gun ownership. However, a different mindset, less enriched in learned helplessness, will ask what other strategies can be tried. Especially for domestic terrorist-style mass shootings, the answer is obvious: surveillance powered by big data, the increasingly progressive role in our world seems impossible. prevented in all cases.
Information exists, its almost instantaneous aggregation into an understandable pattern after the killing of Buffalo and so many other people testifying. In a twist of irony, the New York Times on Tuesday ran a series of lamentations about the causes of gun control, while another lamented the irresistible pervasiveness of the workplace. Employee monitoring software. As the paper explains, “corporate employers’ fear that employees may leak information, allow access to confidential files, inappropriately or extreme contact with customers are bring a gun to the office.”
Because data exists, because surveillance is cheap, because not doing so puts the business and its public at risk, employers tend to try to solve problems by looking Samples are available for detection: “The software can track suspicious computer behavior or it can dig into employee credit reports, arrest records and marital status updates. . It can check if Cheryl is downloading bulk cloud data or run sentiment analysis on Tom’s emails to see if he’s received testimonials over time. ”
Longtime readers will sigh. I made similar views after half a dozen domestic terrorism-style events from Las Vegas and suburban Denver to a congressional ballpark in upstate D.C. Red flags, police calls, and suggestions electronics and gifts are always conspicuous in hindsight. A decade ago, it could have been argued that algorithms would have been too slow to come up with suitable samples and would have produced too many false positives. However, throwing away a decade is hardly the way to make progress on these challenges.
The real obstacle is the privacy risk. The privacy risk, let us realize, lies in who can see the data, not whether it exists, when and how it can be allowed to bind an important model. potential importance to a named individual.
In 2017, researchers from Columbia University and Microsoft showed that individual search engine users’ queries can yield recognizable patterns that connect non-specific symptom searches with Subsequent searches showed the user had received a diagnosis of pancreatic cancer. Of course, because the researchers couldn’t identify the individuals involved, they couldn’t ask if anyone had actually received such a diagnosis, and they couldn’t put their insight into helping. Patients who actually get an earlier diagnosis of a disease are often diagnosed too late. Help.
It’s a privacy barrier. A reasonable solution would be to wrap the whole puzzle in a specialized legal process: Algorithms would be allowed to do their job; A judge’s permission will be required before a named person can be linked to an observable pattern so that government officials can take steps. Opportunity exists whether we choose to take advantage of it or not, but history shows that sooner or later we will take advantage of it.
Copyright © 2022 Dow Jones & Company, Inc. All rights reserved. 87990cbe856818d5eddac44c7b1cdeb8
https://www.wsj.com/articles/massacre-data-artifical-intelligence-buffalo-shooter-tops-supermarket-payton-gendron-name-11652817738 Massacre Data Arrives a Day Late