Friday, November 22, 2024

AI is being used to send some households impacted by Helene and Milton $1,000 cash relief payments

Must read

Nearly 1,000 hurricane-impacted households in North Carolina and Florida will benefit this week from a new disaster aid program that employs a model not commonly used by philanthropy in the United States: Giving people rapid, direct cash payments.

The nonprofit GiveDirectly plans to send payments of $1,000 on Friday to some households impacted by Hurricanes Helene and Milton. The organization harnesses a Google-developed artificial intelligence tool to pinpoint areas with high concentrations of poverty and storm damage. On Tuesday, it invited people in those areas to enroll in the program through a smartphone app used to manage SNAP and other government benefits. Donations will then be deposited through the app’s debit card.

The approach is meant to deliver aid “in as streamlined and dignified a way as possible,” said Laura Keen, a senior program manager at GiveDirectly. It removes much of the burden of applying, and is intended to empower people to decide for themselves what their most pressing needs are.

It won’t capture everyone who needs help — but GiveDirectly hopes the program can be a model that makes disaster aid faster and more effective. “We’re always trying to grow the share of disaster response that is delivered as cash, whether that is by FEMA or private actors,” said Keen.

The influx of clothing, blankets, and food that typically arrive after a disaster can fill real needs, but in-kind donations can’t cover getting a hotel room during an evacuation, or childcare while schools are closed.

“There is an elegance to cash that allows individuals in these types of circumstances to resolve their unique needs, which are sure to be very different from the needs of their neighbors,” said Keen. She added that getting money into people’s hands fast can protect them from predatory lending and curb credit card debt.

The organization employs direct payments for poverty relief around the world, but it first experimented with cash disaster payments in the U.S. in 2017, when it gave money to households impacted by Hurricane Harvey in Texas and Hurricane Maria in Puerto Rico. Back then, GiveDirectly enrolled people in person and handed out debit cards activated later. The process took a few weeks.

Now that work is done in days — remotely. A Google team uses its SKAI machine-based learning tool to narrow down the worst-hit areas by comparing pre- and post-disaster aerial imagery. GiveDirectly uses another Google-developed tool to compare those findings with poverty data. It sends the target areas to Propel, an electronic benefits transfers app, which invites users in those places to enroll.

Latest article