{"id":206305,"date":"2020-02-21T01:26:29","date_gmt":"2020-02-20T22:26:29","guid":{"rendered":"http:\/\/ww-vb.mine.nu\/w108\/google-drops-gender-labels-from-image-recognition-to-reduce\/"},"modified":"2020-02-21T01:26:29","modified_gmt":"2020-02-20T22:26:29","slug":"google-drops-gender-labels-from-image-recognition-to-reduce","status":"publish","type":"post","link":"https:\/\/hameed.nwar.uk\/sa\/google-drops-gender-labels-from-image-recognition-to-reduce\/","title":{"rendered":"Google drops gender labels from image recognition to reduce"},"content":{"rendered":"

[ad_1]
\n
<\/p>\n

\n

Google will no longer identify people by gender in its image recognition AI, by removing labels such as \u201cman\u201d and \u201cwoman\u201d from photos of people.\u00a0Instead, every individual will now be classified as a \u201cperson,\u201d\u00a0 according to a company email seen by Business Insider.<\/p>\n

The changes will be introduced to Google Cloud\u2019s Vision API, which developers can use to\u00a0add labels to images and then classify them into predefined categories.\u00a0<\/span><\/p>\n

In an email to developers, Google cited two reasons for the changes: it\u2019s impossible to infer someone\u2019s gender by appearance, and attempting to do so could perpetuate unfair biases.<\/p>\n

Journalist Sriram Sharma shared a screenshot of the email:<\/a><\/p>\n

\n

Got an email from Google saying Cloud Vision API will not return gendered labels such as ‘man’ and ‘woman’ after February 19, 2020. pic.twitter.com\/9XjdgQQwNe<\/a><\/p>\n

\u2014 Sriram Sharma (@SriramVSharma) February 20, 2020<\/a><\/p>\n<\/blockquote>\n

Google added that removing the labels aligned with the second of its\u00a0Artificial Intelligence Principles at Google:\u00a0Avoid creating or reinforcing unfair bias.<\/p>\n

Image recognition systems have a unique tendency to do this.<\/p>\n

In one study, researchers found that\u00a0algorithms\u00a0trained on a deliberately-biased dataset of cooking-related images, in which women were 33% more likely to appear, became 68% more likely to predict a woman was cooking \u2014 even when the image was of a balding man in a kitchen.\u00a0Image recognition systems also regularly misgender trans and non-binary people<\/p>\n

[Read:\u00a0Automated facial recognition breaches GDPR, says EU digital chief]<\/em><\/p>\n

Not everyone will agree with Google\u2019s decision to remove gendered labels from images. Business Insider notes that one developer\u00a0accused Google of prioritizing political correctness over product quality.<\/p>\n

But the move will at least reduce one area of AI bias.<\/p>\n

As linguist and programmer Angus B. Grieve-Smith\u00a0explained on Twitter<\/a>: \u201cAnytime you automatically classify people, whether that\u2019s their gender, or their sexual orientation, you need to decide on which categories you use in the first place \u2014 and this comes with lots of assumptions.\u201d<\/span><\/p>\n

\n

You\u2019re here because you want to learn more about artificial intelligence. So do we. So this summer, we\u2019re bringing Neural to <\/span>TNW Conference 2020,<\/span> where we will host a vibrant program dedicated exclusively to AI. With keynotes by experts from companies like Spotify and RSA, our Neural track will take a deep dive into new innovations, ethical problems, and how AI can transform businesses. <\/span>Get your early bird ticket<\/span> and <\/span>check out the full Neural track<\/span>.<\/span><\/em><\/p>\n

\n Published February 20, 2020 \u2014 17:34 UTC\n <\/p>\n<\/p><\/div>\n