Discover more from PATENT DROP
PATENT DROP: Visa’s AI spending temptress
Plus: Nvidia brings AI into doctor's calls; Ford is all-in on electrification
Happy Monday and welcome to Patent Drop!
Today, we're checking out Visa’s goal to hold onto its customers using AI, Nvidia’s tech to bring neural networks into the doctor’s office, and Ford’s plan to take on another kind of electrification.
Here we go.
#1. Visa’s AI persuasion check
As a credit card company, it’s in Visa’s best interest to keep its customers spending. AI might be able to help.
The company seeks to patent tech that applies deep learning to analyzing “financial device usage.” This tech uses an AI model trained on historic transaction data, monitoring your spending activity and how often you use your “financial device,” a.k.a. your credit or debit card, to predict the probability of that card becoming – or ceasing to be – your primary card.
With this information, Visa can then send you targeted rewards to influence your decision to drain our account or not.
“If an issuer can anticipate when a consumer will change their primary financial device, the issuer can offer rewards or perks to ensure the change occurs (if the current primary financial device is not associated with the issuer) or prevent the change from occurring,” Visa said in its filing.
For instance, say you normally use a Visa credit card to pay your rent, buy your coffee or book your plane tickets, but are approved for an American Express card that can get you cash back, points, or other rewards for making those same purchases. With this tech, Visa might fight back and try to tempt you into using your card by flooding your inbox with travel rewards or cash-back deals.
We’ve seen tons of examples of AI making its way into the fintech space, from PayPal using machine learning to track down fraud to Mastercard creating generative neural networks to prevent phishing. And plenty of financial institutions use run-of-the-mill AI digital assistants for customer service. But Visa’s patent is turning AI into its salesmen by training it to influence customer acquisition and retention, and has the potential to “revolutionize” the field by more accurately predicting a customer’s behaviors, said Iu Ayala Portella, CEO and founder of machine learning consultancy Gradient Insight.
“Imagine being able to anticipate a customer's financial needs before they even occur,” Portella told me. “With the use of deep learning analysis, this is now possible.”
This tech follows the same playbook of almost every AI-based tech innovation today: It watches you and connects the dots to predict your patterns and use them in some way, shape, or form, said Jake Maymar, VP of Innovation at The Glimpse Group. A financial institution like Visa already has access to hoards of data on its customers. With this patent, the company is simply putting that data to use.
“The more data that's available, the more it's easy to sift through and actually see these patterns,” said Maymar.
As AI develops, Visa’s tech is far from the last time we’ll see innovations like this pop up among financial companies, Maymar said. With the sheer amount of data that financial companies have from millions of individual consumers, the possibilities for training AI models are practically endless. What’s next is still unknown, he said.
“If I'm a credit card company and I have all this data ... What other information can I glean from that? That's really the big question,” said Maymar. “They’re filing a patent so that they can keep customers happy, but what else are they going to do with this power?”
#2. Nvidia, the physician’s assistant
Nvidia wants to get between you and your doctor.
The company filed a patent application for tech that estimates heart rate and respiratory rate using neural networks with higher accuracy and lower latency “compared with conventional systems.” Here’s how it works: this tech takes image data of a subject to estimate vital signs using a motion neural network, or an AI which uses a “motion map” and a sequence of images to track subtle movements.
This neural network focuses specifically on facial features that have a “strong physiological signal response,” rather than the whole picture, to tune out the noise and produce results with “higher accuracy, better precision, and greater efficiency than otherwise possible with conventional systems.”
Nvidia noted in the patent that this tech is helpful in telehealth contexts, as it provides a doctor with an accurate way to measure vital signs without actually needing to be in the room with the patient. This could be helpful in cases of contagious disease… like a global pandemic.
“Non-contact camera based physiological measurements tend to be more accessible and convenient in Telehealth,” Nvidia said in its application. “However, remote physiological signal measurement is challenging due to environment illumination variations, head motion, facial expression, etc.”
Nvidia’s patent is just one example of the impact that AI could have in the medical sector. Whether it be something as simple as automating note-taking, and record-keeping, or as personal as patient diagnoses and tracking your health in real-time, AI is vastly altering healthcare at the same clip as seemingly every other industry.
Nvidia bridging its AI talents with healthcare and medical tech makes perfect sense. The company has a tight grip on the chips business over its competitors like Intel and AMD, and that dominance could position it to be one of the biggest frontrunners in AI chips and development.
Nvidia also has a healthcare division that works on implementing its computing capabilities into things like medical imaging, biopharma, and genomics, though this business unit is far from its biggest money maker. However, medical technology is an incredibly lucrative – and largely recession-proof – sector of tech: An October report by Market Research Future estimated that the global medical technology market could be valued at $964 billion by 2030.
Nvidia makes a large chunk of its money from selling chips for consumer computers and devices. The company strengthening its medical tech unit could provide it with a solid paycheck in rough economic waters, as its consumer unit could face hardship in the short term amid a potential recession.
#3. Ford’s electric feel
Ford is plowing forward with its electrification goals, and using hydrogen power to get there.
The auto manufacturer filed a patent application for a “hydrogen fuel tank mounting system.” Basically, this system mounts a hydrogen fuel tank under the vehicle, with mechanisms in place that allow it to expand and contract, taking into account that this kind of fuel tank often expands while being filled and contracts while being used.
Ford’s tech allows for a lighter-weight vehicle, such as a standard car, to hold a much larger hydrogen fuel tank. Other methods, such as keeping the fuel tank within the body of the car, limit “the potential size and weight of the storage tank,” ergo, limiting how far a car can travel before needing to refuel.
As trivial as this invention sounds, Ford noted that this system is essential to reduce stress on other parts of the vehicle and improve its durability.
“Hydrogen as a fuel for automobiles and other transportation vehicles is expected to increase in popularity in the near future. Onboard storage of hydrogen presents a number of challenges,” Ford said in its filing. “These challenges may be particularly marked in the storage of hydrogen in light vehicles, such as automobiles.”
Essentially, in a hydrogen fuel cell vehicle, size matters: The bigger the tank, the farther it can go, as energy output is determined by tank size. But Ford’s patent might provide a solution in helping these vehicles to go the distance.
It’s no secret that Ford is all-in on battery-electric vehicles. The company has ambitious goals for half of all vehicles it sells to be electric by 2030 and has continued to apply for patents for EV-related inventions, including an EV charging reservation system and carbon footprint estimation tech that we covered in Patent Drop back in January.
And while traditional EVs that rely on lithium ion batteries are the current focus of many carmakers, hydrogen fuel cell-powered vehicles are another alternative to traditional gas cars with major potential. According to the Department of Energy, fuel cells don’t emit pollutants from their tailpipes – just water. Plus, as mining for critical minerals becomes more competitive with increasing demand for battery EVs, hydrogen fuel cell vehicles provide a clean energy alternative with an already abundant fuel source.
As it stands, the market for these cars is currently minuscule. Only three companies offer hydrogen fuel cell-based vehicles –Toyota, Hyundai, and Honda – with around 15,000 of these EVs on the road in the U.S. as of mid-2022, according to Car and Driver. Ford might be trying to break into the game while these cars are still under the radar.
The company is already reportedly working on a utility-grade vehicle powered by hydrogen fuel cells. And if this patent is any indication, it shows that Ford is going after mass commercial electrification of its fleet in more ways than one.
We’ve got a few more to show off.
Want to dress snazzy in the metaverse? Meta’s working on it. The company is seeking to patent “digital garment generation” that allows users to take photos of their favorite clothes and translate them into a virtual object, to make their avatar reflect their personal style.
Boeing is taking notes on Amazon’s delivery drones. Insitu, a Boeing subsidiary, filed a patent for “package delivery systems” involving drones. Though this sounds like it’s been done before, these would likely be used in a military context, not a consumer one.
Google is making a virtual mime. The company filed a patent for a “Puppeteering Remote Avatar” with facial expressions, to “reflect a user's personality and character” in a video call.
What else is new?
Baidu debuted some of the capabilities of its AI chatbot, Ernie. The platform can summarize financial documents, produce PowerPoints, create travel itineraries and more.
Neuralink is getting ready to start testing on humans. The company approached a major U.S. neurosurgery centers as a potential clinical trials partner, Reuters reported.
Parts of Twitter’s source code were leaked on GitHub without the app’s permission, according to a court filing. GitHub has since taken down the code.
Interested in how policy and the trajectory of the economy are dictated by the collision of money and power? Or how special interest groups shape our society? Introducing Power Corridor, the latest enterprise newsletter from The Daily Upside. Power Corridor aims to take you behind closed doors and into the room where Wall Street and Washington overlap. From big themes in the news cycle to changing global power balances, Power Corridor will break down the key players, events, and ideas shaping the economic terrain that lies ahead. Get investigative reporting and analysis directly to your inbox — sign up here