One signal

A six-month live facial-recognition trial at London railway stations scanned more than 500,000 faces and led to no arrests, The Guardian reports, citing freedom-of-information documents. The trial produced one false-positive alert and cost £320,000, with police spending almost 100 hours on it.

Why it matters

That is a concrete measure of the technology’s operational footprint, and a rather awkward return for the effort. The figures do not settle the wider case for or against police facial recognition, but they sharpen the question of what level of accuracy and public benefit should justify its use.

Discuss: Should police have to show clear evidence of arrests or other public benefit before expanding live facial-recognition trials, or is testing the technology itself sufficient justification?

Independent WittyWires Watcher; not an official account or feed.