South Africa is bringing machine learning to its border checks

Phathisani Moyo


South Africa is turning to machine learning, biometrics and facial recognition to tackle one of its most politically explosive problems: controlling who enters the country.

President Cyril Ramaphosa will launch the Electronic Travel Authorisation (ETA), a digital visa system, at OR Tambo International Airport on Wednesday, making it the centrepiece of Home Affairs’ technology immigration overhaul. 

The move comes as anger over undocumented migration, porous borders and weak enforcement has fuelled anti-immigration protests, some of which have turned violent and strained relations with countries including Ghana and Nigeria.

“The ETA combines advanced biometric verification, machine learning and the upgraded Electronic Movement Control System (eMCS 2.0) as part of a modern digital immigration ecosystem that strengthens border security while making travel to South Africa faster, simpler and more secure for legitimate travellers,” the Presidency said in an August 6 statement.

The system has already processed more than 203,000 applications since May before it officially launches on Wednesday, according to Leon Schreiber, the Home Affairs Minister, with more than 5,500 fraudulent applications rejected, including fraudulent passports, manipulated documents and other indicators of fraud detected through machine learning.

In his May budget vote, Schreiber said the ETA also allows prospective travellers to apply for a tourist visa from their laptop or smartphone, using biometric and machine learning technology to verify their identity.

The technology starts checking travellers before they reach the border. Schreiber said the system checks 40 parameters to establish whether a passport is authentic and uses liveness detection to compare an applicant’s selfie with their passport photograph. At the…



Source link
Read Full Article by Phathisani Moyo at techcabal.com
Source link

Share This Article
Leave a Comment
Home
Account
Shop
Community
Add. A Post