DHS biometrics collected from migrant kids could solve FRT training dataset shortfalls | Biometric Update

The US Department of Homeland Security plans to use photos of migrant children to train facial recognition technology, aiming to identify individuals as they age. This initiative involves collecting craniofacial data from thousands of unaccompanied minors, creating a dataset far larger than existing aging research sets. While officials claim this improves identity services, it effectively turns vulnerable populations into test subjects for government algorithms. This practice raises severe ethical and legal concerns regarding consent, given the high-pressure border environment and inherent power imbalances. Civil rights advocates argue that meaningful consent is impossible in such contexts, potentially establishing a surveillance state. Furthermore, relying on real-world data from marginalized groups risks embedding structural biases into algorithms, which could lead to discriminatory outcomes and over-policing of people of color if errors or skewed representations occur. The situation highlights a critical issue for open data: borders act as laboratories for tech experimentation with minimal oversight. Technologies introduced for research purposes often expand beyond their original scope, influencing broader military and global data collection efforts. This case underscores the necessity for strict transparency, ethical guidelines, and public accountability in government data collection, particularly when involving vulnerable communities and sensitive biometric information.

Source: biometricupdate.com
Published on 2024-08-16