DIANE GREENE, Google’s director of cloud computing, has warned that facial recognition lacks diversity and is still inflicted with “inherent biases”.
Greene’s comments, made to the BBC during Google’s Next conference in San Francisco, came shortly after a test of Amazon’s Rekogniton technology by the American Civil Liberties Union. The software incorrectly identified 28 members of Congress – disproportionately those of colour – as people who might have been arrested for a crime (Amazon has said that the ACLU used the wrong settings).
Although Google is working hard on gathering data to improve the reliability of its own service, Greene would not talk about the company’s work with the US military, telling the Beeb: “Bad things happen when I talk about Maven.”
Face recognition has been a major cause for concern amongst the Silicon Valley elite, as well as civil rights groups. In April, Forbes uncovered Face-Int, which pulls data from Facebook to build a massive facial recognition database, while in June Amazon workers called on CEO Jeff Bezos to stop working with law enforcement.
Jay Stanley of the ACLU said of Face-Int: “If private companies are scraping photos and combining them with personal info in order to make judgements about people – are you a terrorist, or how likely are you to be a shoplifter or anything in between – then it exposes everyone to the risk of being misidentified, or correctly identified and being misjudged.”
Greene told the BBC: “We need to be really careful about how we use this kind of technology.
“We’re thinking really deeply. The humanistic side of AI – it doesn’t have the diversity it needs and the data itself will have some inherent biases, so everybody’s working to understand that.
“I think everybody wants to do the right thing. I’m sure Amazon wants to do the right thing too. But it’s a new technology, it’s a very powerful technology.”
Following the ACLU’s test, two members of Congress have written to Bezos about the alleged issue with the technology.
Source : Inquirer