Facial-recognition software inaccurate in 98% of cases, report finds

posted onMay 14, 2018
by l33tdawg
https://www.evolving-science.com/sites/default/files/field/image/Facial-Recognition.jpg
Credit: https://www.evolving-science.com/sites/default/files/field/image/Facial-Recognition.jpg

Facial recognition may not be the high-tech policing solution it's purported to be, with new figures showing facial-recognition software used by the UK's Metropolitan Police returned incorrect matches in 98 percent of cases.

According to figures published by The Independent (based on data obtained under freedom of information), only two of the 104 alerts generated by the facial-recognition software used by Met Police were found to be accurate matches.

The Independent also reports software used by South Wales Police has returned more than 2,400 false positives since June 2017.

Source

Tags

Industry News Artificial Intelligence

You May Also Like

Other Articles In This Section

Recent News

Friday, November 1st

Youth of today say passwords are old news, passkeys are the future
Chinese attackers accessed Canadian government networks – for five years
OpenAI launches ChatGPT with Search, taking Google head-on
Here’s the paper no one read before declaring the demise of modern cryptography
Not just ChatGPT anymore: Perplexity and Anthropic’s Claude get desktop apps

Tuesday, July 9th

China's APT40 gang is ready to attack vulns within hours or days of public release
AI-Powered Super Soldiers Are More Than Just a Pipe Dream
The president ordered a board to probe a massive Russian cyberattack. It never did.
Massive car dealer ransom attack is mostly over after 2 weeks of work-arounds

Wednesday, July 3rd

“RegreSSHion” vulnerability in OpenSSH gives attackers root on Linux
Two of the German military’s new spy satellites appear to have failed in orbit

Friday, June 28th

Indonesian Airports, Data Centres Hit By Worst Cyberattack in Years
Cisco Talos warns of wider security implications following Snowflake breach

Thursday, June 27th

I Wore Meta Ray-Bans in Montreal to Test Their AI Translation Skills. It Did Not Go Well
Researchers upend AI status quo by eliminating matrix multiplication in LLMs
YouTube tries convincing record labels to license music for AI song generator
Critical MOVEit vulnerability puts huge swaths of the Internet at severe risk
Julian Assange to plead guilty but is going home after long extradition fight

Thursday, June 13th

Hackers show off jailbroken checkm8-vulnerable iPad and Apple TV running iPadOS 18 & tvOS 18 respectively
One of the major sellers of detailed driver behavioral data is shutting down
Turkish student creates custom AI device for cheating university exam, gets arrested

Wednesday, June 12th

Chinese-Made Biometric Access System Has 24 Vulnerabilities
Apple and OpenAI currently have the most misunderstood partnership in tech
Adobe to update vague AI terms after users threaten to cancel subscriptions
China state hackers infected 20,000 Fortinet VPNs

Tuesday, June 11th

Russia Is Targeting Germany With Fake Information as Europe Votes
Apple announces macOS 15 Sequoia with window tiling, iPhone mirroring, and more
Apple integrates ChatGPT into Siri, iOS, and macOS
British police to scan 480,000 Formula 1 fans’ faces at Silverstone
Hackers steal “significant volume” of data from hundreds of Snowflake customers

Friday, June 7th

7,000 LockBit decryption keys now in the hands of the FBI, offering victims hope
FCC pushes ISPs to fix security flaws in Internet routing
Can a technology called RAG keep AI models from making stuff up?
How to Lead an Army of Digital Sleuths in the Age of AI

Thursday, June 6th

Mystery object waits nearly an hour between radio bursts