Skip to Content

Why Self-Driving Cars Need ‘Morality’ Settings: Eye on A.I.

Autonomous cars should come with "morality" settings that let car owners control whether their vehicles speed and pass others or follow the speed limit.

Omoju Miller, a machine learning engineer at business software firm GitHub, recommended a morality setting last week during a discussion about the ethics of artificial intelligence at Fortune's Brainstorm Tech conference in Aspen, Colo. The idea is to give humans a direct say in operating self-driving cars and the sometimes life-and-death decisions that the technology powering those vehicles will inevitably make.

For instance, someone who enjoys speeding and cutting off other motorists could choose a more "aggressive" setting. But if that car collides with another, that vehicle's owner would be liable.

“You’re going to pick what morality mode you subscribe to,” Miller said. “The car will make a decision based on your morality.”

Expect more far-out ideas as self-driving technology matures. Currently, there are no specific laws that lay out who (or what) would be liable—the owner, car maker, or technology provider—if a self-driving car collided with another.

****************

In reporting my recent magazine story about "cloud gaming," or streaming video games to smartphones or laptops, I chatted with Ubisoft executive Chris Early about his hopes for A.I.

Developing video games for "cloud" data centers gives game publishers unlimited access to computing power and A.I. technologies that could help them create more realistic games, Early said. For example, many of today's video games feature characters, known as NPCs, or non-playing characters, that are controlled by the game. Usually, these characters are usually fairly "dumb" compared to ones directed by human players.

But machine-learning technology could create "smarter" characters that respond more realistically. As Early said, "Maybe they remember you and treat you differently."

Jonathan Vanian
@JonathanVanian
jonathan.vanian@fortune.com

EYE ON A.I. NEWS

It costs a lot to be “open.” Microsoft is investing $1 billion in A.I. research startup OpenAI, which has previously received financial backing from tech executives like Tesla CEO Elon Musk and LinkedIn co-founder Reid Hoffman, Bloomberg News reported.  The partnership will focus on artificial general intelligence, the point at which computers can learn new skills and complete tasks as well as humans.

Orlando waves goodbye to facial-recognition tech. The Orlando Weekly reported that the city of Orlando is ending its trial of Amazon’s Rekognition software that is used to analyze people’s faces, among other tasks. “At this time, the city was not able to dedicate the resources to the pilot to enable us to make any noticeable progress toward completing the needed configuration and testing,” Orlando’s Chief Administrative Office said, according to the report. 

Arriving on time. Aviation Today reported that Fraport, a Germany-based airport operator, is using machine-learning technology to help predict when airplanes will arrive at the Frankfurt Airport in Germany. The report said that the “predictive aircraft landing time technology uses machine learning models to predict when a flight will actually touchdown on one of Frankfurt’s runways.”

Robot IPOs. CloudMinds, a Beijing-based startup that specializes in A.I. software for robots, plans to raise $500 million as part of a planned IPO on the New York Stock Exchange, reported MarketWatch. The startup had $121 million in sales last year, but lost $156.4 million, the report said.

A.I. IS EASY TO ABUSE

The Scientist interviewed prominent microbiologist Nick Loman about the use of machine learning to aid scientific research. Loman explains that machine learning technologies are powerful, but that they’re also easy to abuse. He says: “They will take in your data set and they will generate a model, but they don’t necessarily tell you that you’ve done the wrong thing.”  In short: “It’s just your classic garbage in, garbage out situation.”

EYE ON A.I. HIRES

The Veterans Affairs Department chose Dr. Gil Alterovitz to be its first director of artificial intelligence, Nextgov reported. Alterovitz is a Harvard Medical School professor and member of the Computational Health Informatics Program at Boston Children’s Hospital.

Satellite television provider Dish Network hired Kannan Alagappan to be its CTO. Alagappan was previously the CTO of telecommunications company Telstra and the senior vice president of customer success and services for enterprise software company New Relic, according to his LinkedIn profile.

Gagosian, a contemporary art gallery owned by Larry Gagosian, hired Sebastian Cwilich to be a part-time senior advisor for the gallery’s technology department, ARTnews reported. Cwilich was previously the co-founder, president, and COO of New York-based Artsy.

EYE ON A.I. RESEARCH

Machine learning takes on climate change. Researchers from Harvard University, MIT, Google, Microsoft, and other institutions published a paper about using machine learning techniques to tackle a variety of problems that could exacerbate climate change. National Geographic noted, “The paper offers up 13 areas where machine learning can be deployed, including energy production, CO2 removal, education, solar geoengineering, and finance.” Although the paper raises hope that A.I. technologies could prevent or slow-down environmental catastrophes, National Geographic cautioned that it’s still “early days and AI can’t solve everything.”

FORTUNE ON A.I.

An Algorithm May Decide Your Next Pay Raise – By Anne Fisher

A.I.’s Hidden Biases Are Continuing to Bedevil Businesses. Can They Be Stopped? – By Jeremy Kahn

A.I. Uses Expected to Expand as U.S. Consumers Warm Up to Trading Data for Convenience – By Danielle Abril

BRAIN FOOD

Ex-Apple chief chimes in on possible Google A.I. plans. John Sculley, a former CEO of Apple, wrote in Fortune about the revolutionary changes that data-collecting sensors could bring to the health-care industry. In the essay, Sculley outlined a possible scenario involving Google: “Google could put advanced medical sensors in its Pixel smartphones, along with its A.I. voice assistant technology that interprets personalized sensor-generated medical data. The company could create a smart, automated patient caregiver for chronic care patients at home. And Google could merge its sensors, along with A.I.-powered Duplex and smart home system Nest, into a health subscription service.”