Amazon’s facial recognition tech has been under controversy for quite some time now. There are a mixture of opinions about the patent with some people having no problem whatsoever with it but with some raising concerns over it when Amazon revealed they would sell the facial recognition tech to government agencies. There are others whose concerns lay with the technology resulting in false arrests seen as the AI was found unable to identify the gender of individuals with darker skin.
None of the above issues seemed to phase shareholders though as shown in a regulatory filing on Friday. Shareholders overwhelmingly rejected a proposal that suggested the company should stop selling their facial recognition technology to government agencies. Only 2.4% of votes were in favour of the ban whilst the remaining votes were voted against.
A second proposal which called for the inspection of how much Amazon’s “Rekognition” service harmed privacy and civil rights gathered 27.5% of votes which is far greater than the amount of votes the ban got. It still isn’t enough for anything to come of it, however, which is a shame.
The facial recognition technology has already been sold to law enforcement in Florida and Oregon which was the spark for the highly controversial debate. Amazon has defended their work by saying all users of the tech must abide by law and not misuse the technology.
Something to note about the votes that took place is that founder and chief executive, Jeff Bezos, had 16% of the stock and voting rights.
After the intense scrutiny Amazon’s facial recognition technology has received U.S. Congress members are looking into the impact the technology would have on rights. So despite the shareholders’ negative reaction to an inspection of the tech it seems it may be happening anyway.
The development of this kind of AI has caught the attention of other vendors as well. France’s Idemia, Japan’s NEC Corp and more recently Israel’s AnyVision and Microsoft Corp which has called for regulation in recent months are all interested in the integration of facial recognition with their own implementations.