As far as facial recognition databases go, the one native to the Federal Bureau of Investigation is getting quite crowded. Reported on by Georgetown Law this week, the FBI’s burgeoning database apparently houses more than 117 million American faces, meaning it currently boasts about half the U.S. population of adults. Don’t break out your tinfoil hat (or mask?) just yet, though, the FBI’s official response shouldn’t necessarily ruffle many feathers — it’s allegedly been compiling the database since 2011 as an effort to assist the criminal justice community.
However, the bureau’s reasoning didn’t stop Georgetown from highlighting potential dangers, including the impact on free speech, personal privacy, and the ability to avoid unreasonable searches and seizures which may stem from access to such a database. According to the 150-page report, roughly 25 percent of state and local police outfits have access to a facial recognition database capable of scanning for driver’s licenses or specific IDs. Perhaps the most startling point made by Georgetown’s report is the fact many of these troves of data are largely unregulated and, as the report put it, “out of control.”
Georgetown Law’s report was the work of Electronic Frontier Foundation — a nonprofit dedicated to “defending civil liberties in the digital world” — which directed attention to the FBI’s work a few years after it began.
Back in 2014, the EFF filed a Freedom of Information Act lawsuit urging the FBI to release data on its as-of-then unprecedented push to establish a facial recognition database. Staring down the barrel of a lawsuit, the FBI obliged the EFF and released a slew of documents depicting its work on an initiative called Next Generation Identification — or NGI.
An advanced identification network built on the back of the FBI’s current fingerprint database — which boasted more than 100 million records as of 2011 — NGI aimed to include additional biometric data such as iris scans, palm prints, and (our feature in question) facial recognition. Once this data is compiled, NGI associates it with someone’s name, address, age, race, driver’s license number, and more. During EFF’s original inquiry, it was expected that this facial recognition database would contain roughly 52 million faces by the end of 2015.
It’s clear NGI not only hit this lofty goal but has wildly eclipsed it in the years since.
As of now, the FBI has legal clearance to use its software to search for driver’s license information in 26 states. Of course, the hope is that the bureau — or any law enforcement agency with database access — resorts to merely using this futuristic capability in specific times of need such as to help with a criminal case or track someone actually worthy of being tracked.
Despite the good intentions of the FBI, it wouldn’t be the least bit unexpected to see a bit of worried backlash in light of Georgetown’s report. Considering the fact the NGI system originally planned on grouping criminal and non-criminal facial information together — a feature first uncovered during EFF’s FOIA lawsuit — there may be reason to be ill at ease about its use.
San Francisco startup Dimensional advertised for people to walk robot dogs through city streets. The New York Times reported that the job could include carrying a roughly 33-pound robot back to the office if it overheated. A lift home is an awkward requirement for a labor-saving machine.
Keeping the hardware running is only one challenge. Anthropic's robotics tests found that language models mostly failed when controlling joints directly, but did better when directing trained controllers. Giving an AI agent a body doesn't teach it to handle ordinary household objects.
Soon, a mattress protector may be more than a bed cover. Australian medtech company Sleeptite and researchers at RMIT University have received a US patent for FlexiSense, a smart bedding technology that tracks body position and posture to help caregivers identify when older residents may be at risk of falling.
The technology is being developed for Sleeptite’s in-bed monitoring system, REMi, and was recently on display at The MedTech Conference in Boston. A peer-reviewed study in Flexible and Printed Electronics says the sensors can also be made at scale without sacrificing performance or durability.
AI companies spend enormous amounts of time convincing us their technology is safe and secure. But what happens when those same companies become targets of increasingly sophisticated AI-powered attacks? Anthropic, the company behind Claude, is now facing that question. According to a Washington Post report, cybersecurity startup Cathedral discovered vulnerabilities that let its researchers access sensitive information in Anthropic's internal systems.
The researchers used AI agents to identify potential entry points in less than 12 hours. Interestingly, they didn't need to crack Claude itself. Instead, they exploited weaknesses in third-party business applications that Anthropic uses for everyday operations. Anthropic's security testing program disclosed the findings, which allows outside researchers to identify and report vulnerabilities.