ARTIFICIAL INTELLIGENCE: Paying by Smile with Alibaba or by Blinking with Ping An


Chinese commerce is very digital already, far outpacing the US in both nominal and percentage terms. Since almost no mobile payments in 2011, China now sees almost 100 trillion yuan, or $14 trillion USD, in mobile payment transaction volume. This compares to less than $100 billion in the United States -- a 10x difference in adoption of using phones, rather than cards or cash, to pay for things. Further, unlike in the West, the vector of payments intersects much more closely with social identity and networking, which is the platform globally for developing artificial intelligence. Just check your Facebook Newsfeed.

So we give to you implementations of AI for payments in the East. The first is from Alibaba. If the customer has Alipay's app and has enabled facial recognition, a smart vending machine is able to scan your face and associate it with the payment account. We would guess that there is a geolocation element involved as well for two factor authentication, or perhaps just a phone or pin verification. The second example is the newly launched Ping An partnership with Danyang Rural Commercial Bank. The plan is to use facial recognition combined with "blink detection" to authorize a payment. The Bank claims to target 1,000 merchants for the initial pilot of the program. Reminder -- Ping An has built out machine vision capabilities to cut down on time processing insurance claims, and here it is trying to rent it out as a cloud service to other providers.

We end with a few questions. First, if Ping An was able to stand up real machine vision capabilities within a couple of years, what's stopping Visa or Mastercard or JP Morgan from building the same? Why have American finance firms failed to own the AI technology layer and its associated cloud? We think the answer has to do with the role of enterprise tech firms and implementation consultants in the US, which make the default option to out-source rather than in-source such capability. Why build, when you get this from Google for free as part of a cloud deployment? And second, we observe that massive data processing and hosting infrastructure is needed to accurately process image recognition on video for millions of people in real time. Likely, you also need high definition images to pick up blinks and smiles. So let's refresh that 5G network!


Source: Walk the Chat (Charts), Fung Global Retail & Tech (Chart), TechCrunch (Alibaba), MPayPass via CrowdFundInsider (Ping An)