May 17, 2019

Axios Navigate

Joann Muller

Good morning! Thanks for reading. Please share this newsletter and tell your friends they can subscribe here. If you have tips or feedback, just reply to this email.

Today Expert Voices contributor Haywood Marsh explores how artificial intelligence can help tackle insurance challenges for AVs.

1 big thing: AVs need to be better mind readers

Illustration: Sarah Grillo/Axios

Self-driving cars can be programmed to stay in their lane and obey speed limits. What they lack is human intuition — the ability to know what's going on inside someone else's head.

Why it matters: AVs are getting closer to reality but if they're ever going to drive better than people they need to learn how to share the road with other cars, pedestrians and cyclists. Turning them into social creatures with human instincts is arguably the biggest roadblock to autonomy.

Human drivers react to social cues all the time — and not just a wave of the hand or a nod from a pedestrian signaling for them to go ahead. People give off hundreds of signals that others use to understand  their state of mind and respond accordingly.

  • A person standing in a crosswalk looking at their phone is probably distracted, for example, so the driver knows to yield.
  • A group of kids playing around on the sidewalk could tumble into the street, so caution is advised.
  • A motorcyclist at an intersection touches their feet to the ground, so it's probably safe to proceed.

Be smart: AVs need the same instincts so they can recognize, understand, and predict human behavior. But that ability, which comes so naturally to humans, is hard for machines to learn.

Driving the news: Perceptive Automata, a startup whose backers include Toyota and Hyundai, has a unique approach — it's applying behavioral science techniques to machine learning.

  • Instead of training AI models by labeling objects (this is a tree; this is not a tree, for example), Perceptive Automata characterizes the way people understand others' state of mind and then trains its algorithms to recognize and predict human behavior.

How it works:

  • Researchers collect sensor data from vehicles interacting with pedestrians, bicyclists, and other motorists and then chop up the images into smaller slices.
  • Then they blur or cover a portion of each slice and ask groups of people what the depicted person is about to do.
  • They repeat the process hundreds of thousands of times, with a variety of interactions, and use the results to train models that interpret the world the way people do.

Humans don't always agree, of course — predicting human behavior is not as easy as labeling a tree — but capturing that ambiguity is important to program how an AV should behave, CEO Sid Misra says.

Yes, but: That also means self-driving cars could be prone to mistakes, notes Elizabeth Walshe, a cognitive neuroscientist who studies driving behavior at the Children's Hospital of Philadelphia and the University of Pennsylvania.

  • "Humans are not perfect. Are we happy with machines that make mistakes and are not perfect?"

The bottom line: It will take time for people to learn to trust self-driving cars, but understanding what they might be thinking is a good place to start.

2. No driver, no cab

Einride T-Pod. Photo: Einride

A driverless electric truck "[r]esembling the helmet of a Star Wars stormtrooper" began daily freight deliveries on a public road in Sweden this week, according to Reuters.

Why it matters: Swedish developer Einride is not the only company working on driverless trucks, but it's unique because its T-Pod truck doesn't have a driver cabin.


  • The cargo-only format reduces freight operating costs by about 60% versus a diesel truck with a driver, the company's CEO told Axios earlier this year.
  • The T-Pod features level 4 autonomy, but a remote operator can supervise and control up to 10 vehicles at once, Reuters reports.
  • For now, it has permission to make short trips between a warehouse and a terminal on a public road in an industrial area.
  • The company is talking to partners about scaling production and aims to have 200 vehicles in operation by the end of 2020, according to Reuters.
3. Insurance for AVs will rely on sophisticated AI

Illustration: Lazaro Gamio/Axios

The insurance industry is learning how to incorporate AI in various scenarios to determine liability and calculate risk, including AVs where responsibility for incidents could fall on backup drivers, the manufacturer, or software and hardware providers, Haywood Marsh writes for Axios Expert Voices.

Why it matters: Insuring AVs is an unprecedented insurance challenge, as they will generate huge volumes of data from a variety of parts and will be comprised of complex systems that share responsibility in interlocking ways.

AI is already used in catastrophic storm insurance, another insurance challenge that involves unprecedented events and massive amounts of data. Insuring AVs will rely on numerous AI-based tools for data analysis to...

  • Build out a preliminary database.
  • Analyze individual incidents.
  • Identify patterns in order to improve future safety.

Go deeper: Read the full post.

Marsh is general manager of NetClaim, an insurance software subsidiary of NAVEX Global.

4. Driving the conversation

Safety findings: NTSB: Autopilot was in use before Tesla hit semitrailer (Tom Krisher — Associated Press)

  • Why it matters: The Florida crash is the second fatality involving a Tesla on Autopilot that drove under a semi-truck, drawing sharp criticism from safety advocates who say the company isn't doing enough to educate drivers about the technology's limitations.
  • Meanwhile, Tesla CEO Elon Musk has said Tesla will have fully self-driving vehicles on the roads sometime next year.
  • Go deeper: Read NTSB's preliminary report.

Standoff: California regulator threatens Trump with 'extreme' auto rules (Jennifer Dlouhy — Bloomberg)

  • Why it matters: The war between the Trump administration and California's top environmental regulator is escalating.
  • Warning: Air Resources Board chairman Mary Nichols says the state would even consider banning gasoline-powered cars to offset higher pollution that would occur if President Trump rolls back federal vehicle emission and fuel economy standards.

View from Toronto: How to build a driverless car (Jason McBride — Toronto Life)

  • The big picture: For a look at what's happening in AVs outside Silicon Valley, Detroit or Pittsburgh, check out this story about chief scientist Raquel Urtasun and her all-female senior leadership team at Uber Advanced Technologies Group's self-driving transportation lab in Toronto.
5. What I'm driving

From the rear seat, you can control just about everything in the Audi A8. Photo: Audi

This week I am driving the luxurious 2019 Audi A8 L quattro, which is as delightful in the back seat as it is in the driver's seat.

Why it matters: The A8 is Audi's flagship, so it gets the latest technologies that will eventually filter down to other Audi models.

  • Unfortunately, it does not include the much-anticipated Traffic Jam Pilot, Audi's new Level 3 autonomy system.
  • Audi has shelved the system for now because of regulatory challenges in the U.S. and Europe.

This car's technology will still blow you away, though, like the predictive suspension that scans the road ahead to skip over potholes and the 4-wheel steering system that gives this 17-foot-long behemoth the turning radius of a nimble A4.

The interior is serene and inviting, with customized lighting and wood trim on the dash that retracts to reveal hidden air vents.

  • A new infotainment system, featuring 2 large touchscreens, is intuitive and responsive.

Rear-seat passengers can customize their experience using a smartphone-sized remote control that pops out of the rear center console, controlling everything from the deeply reclining seats and heated massaging footrest to the infotainment screens.

The A8 is loaded with driver-assistance technology, even without the Level 3 system. New features include...

  • "Hands-on" traffic jam assist technology, which eases stress in stop-and-go traffic.
  • Intersection assist, which warns drivers of oncoming traffic when their view is obstructed.
  • Emergency assist, which detects if the driver is unresponsive and brings the car to a stop in its lane, turns on the hazard lights, and calls 911.

The bottom line: The A8 is a dream, but dreams can be expensive. The starting price is $83,800, but the loaded version I'm driving is $123,045.

Joann Muller