Walmart is using AI to Detect Checkout Theft in 1,000 stores

It’s very interesting to witness how surveillance capitalism rolls-out. Facial recognition and CCTV are popping up here and there. When the biggest companies roll it out for efficiency, we don’t actually bat an eyelash.

眼看在资本的运作下，监控系统被一步步推向市场，这真是一件“有趣”的事情。不管是在这里还是在那里，人脸识别和闭路电视雨后春笋般层出不穷。所以当这些大型公司为了提高效率而使用监控系统时，我们已经处之漠然。

Walmart is also ushering in shelf-scanning and floor scrubbing robots in other stores.

在其他的商店中，沃尔玛还应用了货架扫描机器人和地板擦洗机器人。

Meanwhile, those pesky AIs are getting into everything. So smile, you’re on candid camera.

同时，那些令人讨厌的AI正在进入生活的方方面面。所以，在摄像头面前，还是保持微笑吧。

Computer Vision Comes to Big Retail

大型零售商所使用的计算机视觉技术

Now in June, 2019, Walmart is using computer vision technology to monitor checkouts and deter potential theft and other causes of shrink in more than 1,000 stores. More surveillance, but for the common good no doubt.

现在是2019年的6月份，为了检测潜在的盗窃行为以及预防其他导致收入降低的原因，沃尔玛在超过1000家商店中应用了一项计算机视觉技术，这一技术可以对结账过程进行监控。虽然监控越来越多，但毫无疑问，这有利于大家的共同利益。

Walmart won’t ask for your permission. Walmart has been surveilling its checkout registers using a computer vision technology called Missed Scan Detection to identify when items move past the scanner without having been scanned.

沃尔玛并不需要你的许可才能部署此系统（因为此系统的监控对象是收银员，译者注）。沃尔玛使用的这项名为“遗漏检测”的计算机视觉技术会对收银员的工作过程进行监控，当商品经过了扫描仪却没有进入结账系统时，该系统可以检测并“报警”。

It might help cut losses, but it raises privacy questions. But what’s privacy in an era of information capitalism, really? What’s freedom on an internet where you have to submit your social media accounts to U.S. immigration?

它可能有助于减少零售商的损失，但是伴随而来的隐私问题又该如何解决？在信息资本主义时代，所谓的隐私究竟是什么呢？当你在现实中需要向美国移民局提交你的社交媒体平台账户时，你在互联网上的自由又是什么？

When you think of Walmart you don’t usually jump to advanced technology, but that’s what’s being used to deter theft and other things at the checkout. Walmart has to become like a technology company to compete with the likes of Amazon. Walmart certainly has been showing its chops with AI and robots of late.

提及沃尔玛的时候，你脑海里可能并不会出现什么高科技的玩意。但是为了预防盗窃行为的发生，沃尔玛开始应用了一些先进的技术。为了能够和亚马逊等巨头进行竞争，沃尔玛必须成为一家科技公司。而且，他们不是已经在推出自己的机器人和人工智能芯片了吗？

Competition Creates New Implementations of AI

竞争促进了人工智能新应用

The goal with the cameras is to detect when items aren’t scanned at checkout. If the cameras see an item hasn’t been scanned, it sends a checkout attendant to provide assistance. This is what we call the human-AI hybrid workplace.

实际上，这些摄像头所要的检测目标是那些没有被扫描入账的商品。如果摄像头检测到一件商品通过了收银台，但是却没有被扫描入账，它就会向服务助理发送确认请求（接收到确认请求后，服务助理观看监控回放来确认是否出现了盗窃行为）。这就是我们所说的人-人工智能混合系统。

The company said that it had made an “investment to ensure the safety of our customers and associates”. No mention to shoppers that it has gone live though, of course.

沃尔玛发表声明称，他们进行了“事前调查，以确保该系统对顾客和员工都是安全的”。但是，他们没有提及的是，该系统已经上线了。

Mission Scan Detection for Anti-Theft

用于防盗的遗漏检测

Missed Scan Detection was designed to help reduce theft and other losses, a problem that cost U.S. retailers up to $47 billion in 2017.

“遗漏检测”技术旨在减少盗窃和其他问题带来的损失。在2017年，这些问题对美国的零售商造成了高达470亿美元的损失。

It sounds a bit like what Microsoft should be selling to SMBs and brick-and-mortar retailers. Microsoft engineers of course are cloning AmazonGo like bundles of tech they can eventually sell to retailers.

这听起来像是微软应该向中小企业和零售商销售的产品。当然了，微软的工程师也正在克隆类似于AlphaGo的产品，以便将来能够进行捆绑销售。

Missed Scan Detection uses cameras to help identify checkout scanning errors and failures. It shows how aspects of AI are being rolled out pretty much automatically. Capitalism will merge in the next thirty years with AI in ways we cannot even anticipate yet in 2019.

遗漏检测技术使用AI加持的摄像头来帮助识别发生在结账过程中的错误或者“故障”。这一案例显示了人工智能方方面面（的应用）是如何被“自动”推出的。在未来的三十年，资本将和AI深度结合，而我们现在根本无法预料将来会发生什么。

Walmart is the biggest retailer in the world. In the two years since the system was deployed, Walmart says it has reduced rates of theft, inventory loss, fraud and scanning errors. Of course it has, implementing AI makes for a more efficient world.

作为世界上最大的零售商，沃尔玛表示，自从系统部署的两年以来，盗窃率、库存损失率、欺诈率和扫描错误率均已经减少。当然了，如果没有效率干嘛要使用它呢？

The Surveillance Approach in Consumerism

来自消费主义的监控

Walmart isn’t the only company using AI to watch over checkouts. The tech Walmart is using in over 1,000 of its stores comes from companies like Ireland-based Everseen. Surveillance, Big Data and predictive analytics are the new weapons against consumers in an advertising centric world.

沃尔玛并不是唯一一家使用监控系统对收银过程进行监控的公司。这项在1000家门店中得到应用的技术源于一家爱尔兰公司——Everseen。在这个广告意味着一切的世界中，监控、大数据和预测分析都是针对消费者的新武器。

If consumerism fosters the implementation of a surveillance state, the jump from loyalty rewards points to social credit systems isn’t as fantastical as you think.

如果消费主义促使了监控国家的出现，那么将忠诚度奖励积分转化为社会信用体系，似乎也并不是那么遥不可及。

The Evolution of Data Capitalism

数据资本主义的演变

Though both Walmart and Amazon have been investing in AI and computer vision technology in their stores, the approaches are very different. While Amazon uses AI to create a cashier-less, seamless shopping experience, Walmart has been implementing technology as a way to boost productivity for its employees.

虽然沃尔玛和亚马逊都在其商店中使用了人工智能和计算机视觉技术，但是两者的目的极不相同。亚马逊想要通过其人工智能为客户提供自动化的、无缝的购物体验；而沃尔玛实施的技术则是为了提高员工的工作效率。

It’s the classic empowerment of people vs. automation of costs scenario. Who do you suppose wins that battle in the long-term?

这就是所谓的人的赋权（提高用户体验）和成本控制自动化（减少成本）之间的战争。长期来看，你觉得谁会赢呢？

The cameras track items rather than people. If an item is spotted being put in a shopping bag before it has been scanned at the checkout, the system can call an employee to “help”. Who‘s to say it’s not matching items with our faces? I’m sure in time it will.

摄像头追踪的目标是商品而不是人。如果在结账的过程中，商品没有经过扫描就被放入了购物袋，那么系统就呼叫服务助理“帮忙”确认。但是谁可以保证说，系统在识别的过程中没有对我们的脸部进行识别匹配呢？我认为这一切（人脸识别的嵌入）只是一个时间问题。

That’s how innovation is trending — for data capture. While the computer vision cameras are helpful for catching thieves who think they can get away with not paying for an item, they’re mostly being used to reduce shrinkage. At the end of the day, however, book keeping is about your end consumer and not just your stock of items.

创新的发展趋势就是如此——一切为了收集数据。虽然通过计算机视觉技术，摄像头可以抓获那些自认为可以不用付钱的小偷，但是这项技术的最终目的还是为了减缓业务收缩的趋势。毕竟，簿记更需要关心（来自）消费者的（收入），而不仅仅是库存记录。