US lawmakers concerned by accuracy of facial recognition

US politicians expressed concerns about the accuracy and growing use of facial recognition software, at a hearing on Wednesday.

The technology is being developed by firms including Amazon and Microsoft and increasingly used by law enforcement worldwide.

Some facial recognition technologies misidentify women and people of colour.

Civil liberties and privacy groups have raised concerns about how the data for these programs is being gathered.

"This is some real-life Black Mirror stuff that we're seeing here," said New York congresswoman Alexandria Ocasio-Cortez, a reference to a science-fiction TV show that explores the dark side of technology.

US lawmakers are working on a proposal for a bill to limit the use of facial recognition.

Some tech experts have raised concerned about how growing facial recognition databases- controlled by governments and private companies - are being used.

"I think we need to pause the technology and let the rest of it catch up," said Meredith Whittaker, co-director of New York University's AI Now Institute and a witness at the hearing.

She argued rules needed to be put in place requiring consent for facial recognition software. Currently, in the US it is enough for a person to be able to see the camera to grant consent.

Ms Whittaker said corporate interest should not be allowed to "race ahead" and incorporate this technology into their systems without safeguards.

Companies that have said they are working on developing facial recognition programs include Google, Microsoft, Amazon and Facebook.

Companies collect data to build their facial recognition software in multiple ways including through CCTV footage and by scraping websites including photo-sharing sites like Flickr and Facebook.

Photo filters like the ones used on Instagram and Snapchat can also be used to hone details of faces.

Police forces in the US have pointed to the successful use of facial recognition technology to identify missing children and criminals.

But biases in the algorithms have led to misidentification. Those accused of crimes because of facial recognition software are often not told the technology has been used.

There are also concerns the technology could be used by authoritarian regimes to monitor citizens.

Countries like China already have extensive surveillance systems set up. Facial recognition makes it easier to track a person's movement.

While the technology is not flawless now, some lawmakers wondered about the implications as the technology becomes more accurate.

"If we only focus on the fact that they're not getting it right with facial recognition, we missed the whole argument," said Rep Mark Meadows, from North Carolina.

"Irrespective of its accuracy, there are intrinsic concerns with this technology and its use," Representative Gerry Connolly from Virginia told the panel.

This was the third hearing the US House of Representative Oversight Committee has held on the facial recognition technology in less than a year.