Improving artificial intelligence to help prevent heart attacks doesn’t just depend on innovation. It also hinges on creating standards, the wonky underpinnings of the technology agreed to by companies and government agencies.
For example, the Consumer Technology Association, which hosts the annual CES tech show in Las Vegas, recently created an A.I. working group made up of gadget makers and healthcare organizations. The goal is to establish best practices for how health-related devices collect, process, and exchange people’s personal data.
It’s not sexy stuff. But standards are critical to making A.I. more useful in nearly every industry.
General Motors, Ford, and Toyota, for instance, recently founded a group to establish standards related to self-driving cars. Meanwhile, the Institute of Electrical and Electronics Engineers, representing industries like telecommunications and electrical engineering, recently formed an umbrella organization for groups working on A.I.-related issues like bias in algorithms, data collection, and autonomous robots.
It’s all a product of A.I.’s relative youth. The technology has emerged so quickly that, in most cases, there is still no consensus within individual industries about how to gather or even interpret data.
Consider fitness trackers that collects heart-rate data and then, potentially, alert doctors when a patient is about to go into cardiac arrest. A lack of agreement about how to precisely measure heart rates could keep the technology from living up to its promise.
Michael Hodgkins, chief medical information officer the American Medical Association, a member of the Consumer Technology Association’s A.I. health group, points to the widespread problem of clunky software in medicine and how it contributes to the high rate of job burnout by healthcare professionals. It’s such a sore spot that The New Yorker recently wrote about the challenge.
“We’ve seen too much technology becoming burdensome,” Hodgkins said recently at a press event.
The Consumer Technology Association’s A.I. healthcare group is also partly intended to convince the federal government to limit regulation of A.I. The association’s CEO, Gary Shapiro, argues that too much regulation may slow innovation.
EYE ON A.I. NEWS
You could have fooled me. Google Duplex, the search giant’s restaurant-booking service that uses a computerized voice that sounds surprisingly human, actually relies on a lot of humans to make phone calls and book tables, according to The New York Times. It turns out that Google needs the human voice-data to further train and improve the A.I. system.
Amazon’s face-scanning plans move forward. Amazon shareholders rejected proposals intended to stop the company’s cloud computing arm from selling facial-recognition technology to law enforcement, USA Today reported. The vote coincided with a House Oversight and Reform Committee hearing about facial recognition technology’s impact on civil rights, in prelude to potential legislation about its use by law enforcement, tech news site VentureBeat reported.
A.I. doesn’t come cheap. Asia-Pacific countries will spend $5.5 billion on A.I. technologies in 2019, an 80% increase from the previous year, according to research firm International Data Corporation. The report said “China will deliver nearly two thirds of the Asia/Pacific excluding Japan (APEJ) regional spending on AI systems in all forecast years.”
If the Mona Lisa could talk. Tech publication Motherboard reported on research by Samsung’s Moscow A.I. center that showed how A.I. could create realistic videos based on a few photos, sometimes even just one. In one wild example, the researchers created a video showing the subject of Leonardo da Vinci’s Mona Lisa painting talking like she’s giving an interview.
TWO BRAINS ARE BETTER THAN ONE
The future of A.I. and healthcare could involve doctors working in tandem with A.I. systems, according to a report by Healthcare IT News. “What we found which consistently (in studies) was that when the doctors took the machines’ suggestions, they were better than either the doctor or machine alone,” said Dr. Ngiam Kee Yuan, the group chief technology officer of the National University Health System in Singapore. “I think that’s what we really want, that is, a combination of AI and a doctor is better than either of them.”
EYE ON A.I. HIRES
Real estate service Zillow Group hired Sing Bing Kang to be its distinguished scientist. Kang was previously a principal researcher at Microsoft’s research arm specializing in computer vision.
WW, formerly known as Weight Watchers, hired Don Kittle to be the company’s vice president of engineering. Kittle, previously a director at marketing firm Loyalty Platform, will help build a data science and engineering team for WW in Toronto.
Pharmaceutical giant Roche picked Igor Korolev to be data scientist for its personalized healthcare analytics team. Korolev was previously the chief scientist of healthcare startup Brainformatika, where was also a co-founder.
EYE ON A.I. RESEARCH
Facebook’s robot projects. Facebook researchers published papers detailing the company’s A.I. work related to robots. The papers cover topics like teaching robots to roll balls or manipulate joysticks through tactile sensation. Check out Fortune’s photo essay that details some of Facebook’s robot studies, as well as other research initiatives from companies like Levi’s and Ford.
FORTUNE ON A.I.
Amazon Prime Boss Named CEO of Google-Backed Quantum Computing Startup – By Robert Hackett
Old Rivals Microsoft and Sony Team Up to Take on New Rivals – By Lisa Marie Segarra
How This NYU Grad Landed an Entry-Level Job at Google – By McKenna Moore
A.I. Captain. Fortune’s Aaron Pressman reports on the experimental autonomous warship Sea Hunter created by defense contractor Leidos. The vessel recently completed a 10-day, 4,000-mile round-trip voyage from Pearl Harbor to San Diego, making it “the first autonomous ship to make an ocean crossing and, remarkably, the first Navy ship designed from scratch by Leidos.” The reports explains that self-navigating ships could help save the Navy money, with one DARPA study suggesting that “Sea Hunter can operate for $20,000 per day, compared with $700,000 to run a fully manned destroyer performing similar mission.”