Amazon.com will release warning cards for products sold by its cloud computing division . The cards will be public so that business customers can see the limitations of such cloud services .
Las Vegas (January 23): In reaction to a persistent question that artificially intelligent systems can discriminate against different classes, Amazon.com Inc is planning to release warning cards for products sold by its cloud computing division, the organization told Reuters.Amazons so-called AI Service Cards will be public so that business customers can see the limitations of such cloud services, such as facial recognition and audio transcription.Amazon said the ultimate goal would be to discourage fraud by using its software, explain how its systems work, and protect personal information.Amazon claimed that the organization is not the first to issue such warnings.
Amazon has also released even more details on the datasets used to train some of its AI, according to the No.3 cloud service provider, Alphabet Incs Google, despite a public outrage on the topic a decade ago when it was assumed that it cared less about AI ethics than its peers.The move would take place at the company's annual cloud conference in Las Vegas, according to Michael Kearns, a University of Pennsylvania professor and Amazon scholar since 2020.The decision to issue the cards came as a result of privacy and fairness checks of the company's software. The most notable feature of this launch is that it promises to do so on an ongoing and an extended basis, he said.Amazon selected software that addresses obscure demographic questions as a start for its service cards, which Kearns hopes to improve in detail with time.Amazon denied a report in 2019 that revealed that the technology was unable to determine the gender of people with darker skin tones.Amazon has extended a moratorium on police use of its facial recognition software after George Floyd, an unarmed Black man, was shot during an arrest in 2020. Amazon also warns against matching faces in cartoons and other nonhuman beings in a recent warning card seen by Reuters.On audio transcription, Amazon claims that changing audio inputs inconsistently could have detrimental effects on different demographic groups.Amazon had worked to address the problem of accurately transcribing the wide variety of regional accents and dialects in North America alone, according to Jessica Newman, director of the AI Security Initiative at the University of California at Berkeley.She said technology giants were increasingly relying on their goodwill to provide the basic information of systems that can have a major effect on peoples lives while still being sufficiently specific and up to date to reflect frequent software updates.