Whether it\u2019s cloud computing or artificial intelligence, the graph of contributions from Amazon has been rising exponentially. Now, the e-commerce giant has announced that one of its pet projects, the facial recognition software Rekognition, can now detect fear.\u00a0\n\nOriginally launched in 2016, Rekognition is Amazon\u2019s facial feature software which adds image and video analysis to the user\u2019s applications. The software had already generated interest (and controversy) for being able to track the facial emotions of a human, like, happiness, sorrow, anger, surprise, disgust, calm and confusion. But now, with the update to recognise fear, Rekognition, has expanded its nucleus for numerous crucial, yet contentious applications.\u00a0\u00a0\u00a0\n\nFear The Fear Itself\u00a0\n\nAccording to a recent official blog, the e-commerce giant has also launched other accuracy and functionality improvements to the face analysis feature in Rekognition. The release is also said to improve the accuracy of gender identity as well as improved age range estimation accuracy.\n\nHowever, this fear recognition update was met with a wave of criticism, by experts claiming that it had a \u201cBig Brother vibe\u201d to it and was \u201ctoo creepy\u201d.\n\nhttps:\/\/twitter.com\/QuinnyPig\/status\/1161058694815072256\n\nBut Amazon has already mentioned that the emotion which the software is detecting is only the physical expression of the face and it is not directly related to the internal emotions. While determining the emotion, Rekognition detects the emotions that appear on the face along with the confidence level.\n\nWide Usage\n\nAccording to the researchers at Amazon, Rekognition provides highly-accurate facial analysis and facial recognition and has the capability to detect, analyse, and compare faces for a wide variety of use cases, including user verification, classifying, people counting, and public safety. It provides benefits such as simple integration of powerful image and video recognition, deep learning-based image and video analysis, scalable image analysis, cost-effective as well as integration with other AWS services.\n\nSome of the most common use cases of this facial detection software are searchable image and video libraries, face-based user verification, sentiment, and demographic analysis, facial recognition, unsafe content detection, text detection, among others. The software can also flag any inappropriate.\n\nControversies\n\nSince its launch, Rekognition has always been surrounded by controversies. It has mainly become the subject of controversy when the law enforcement agencies across the US agreed to use this software. The Washington County Sheriff Office, Oregon is using this tool in their law enforcement department and has claimed to perform well in the process. Law enforcement departments from Arizona and California also showed interest to utilise this tool in criminal investigations. While, according to reports, the American Civil Liberties Union (ACLU) did not agree with the same and provided pieces of evidence that the tool is not fit to take sophisticated decisions. The ACLU compared its members of congress with 25,000 mugshots and criminal where Rekognition detected that 28 of the congress members matches with the criminals from the database.\n\nhttps:\/\/twitter.com\/RepJimmyGomez\/status\/1022496878250868736\n\nSimilar Offerings\n\nGoogle is working on a similar technology called Vision AI which offers several options to integrate computer vision models into a user\u2019s applications and web sites. Vision AI provides AutoML Vision, Vision API and Vision product search. For instance, AutoMl Vision for image classification enables a developer to train custom models that automatically classify images according to the provided labels.\n\nOutlook\n\nAmazon Rekognition is doing far more than just a facial recognition tool. It is being used by the organisations in the marketing as well as advertising space, social platforms to detect various unsafe contents, financial systems for text and sentiment analysis, and much more. However, facial recognition technology is not yet completely developed as it makes misidentification while detecting women and people of colour.