Gaffe of Google Tag it that blacks as ‘gorilla’, then apologizes – AGI – Agenzia Journalistic Italy

Tech

22:14 July 3, 2015

(AGI) – Rome, July 2 – The 2015 will not pass ‘course in history as one of the years most’ rosy history of Google, at least in terms of image. After the flop of ‘Google Glass’, the Mountain View company and’ committed a colossal blunder because the algorithms of Google Photo, which apparently have some problems in distinguishing people from ethnic African gorilla. The case and ‘born two days ago by the complaints of a user, Jacky Alcine, he posted on Twitter a screenshot showing the bizarre way in which Google Photo has organized his photographs in categories:’ skyscrapers ‘,’ airplanes ‘,’ car ‘,’ Bicycle ‘and so far so good. After that
‘ appears a photo of the user, a black, along with her friend, also black, with the inscription above ‘gorilla’.
> “Google Photos, y’all fucked up. My friend’s not a gorilla ‘and’ the text of the tweet. I followed by others ‘tweets’ with which Alcine, a professional developer, showed that, by inserting ‘gorilla’ in search of his library, kept appearing photos of him with her friend. Two hours from the first ‘tweet’ and ‘included in the call Yonatan Zunger,’ Chief Architect of Social ‘of Mountain View, whose team and’ then managed to find a solution to the problem in about an hour. The next morning, unfortunately, two more photos of Alcine and girl appeared labeled as gorillas.
Zunger was therefore unable to do anything but eliminate ‘gorilla’ from the labels of Google Photo, waiting to solve the problem permanently . The company and ‘that “shocked” and apologized to you, promising a thorough investigation into the causes of the error. “There ‘s definitely still a lot of work to do with the automatic labeling of the images and we are trying to understand how to prevent similar mistakes in the future,” said a spokesman. According to the ‘Wall Street Journal’, the incident is linked to the malfunctioning of facial recognition algorithms to Google Photo. “We need to significantly change the systems of machine learning by providing more context, so that ‘they can understand the sensitivity’ cultural importance for humans,” said the Wall Street Journal, the head of the scientific Sentient Technologies, Babk Hodjat, ” humans are very sensitive and focus on specific differences that are culturally relevant to us, the machines can not do. ” (AGI).