As carmakers work to get fully self-driving cars on the road, they have to think about the ethical dilemmas that the vehicles might face. While supporters say the tech will make driving safer, autonomous cars may still have to choose between hitting an animal or swerving into traffic. Brett Berk joins host Zoe Thomas to discuss how companies are grappling with these issues. And Gen Z-ers are supposed to be "digital natives.” So why can so few of them touch-type?
Sign up for the WSJ's free Technology newsletter.
Learn more about your ad choices. Visit megaphone.fm/adchoices
[00:00:00] Hey, you with the podcast in your ear!
[00:00:02] Just a moment!
[00:00:03] Have you already activated your mobile happy hour?
[00:00:05] So you have an unlimited data volume every day for at least one hour.
[00:00:10] It's quite new and for everyone who is now getting mobile phone plus internet for home at Telekom.
[00:00:16] Now Magenta 1 customers will activate the mobile happy hour in the My Magenta app
[00:00:21] and it's already off!
[00:00:22] The Telekom 5G network.
[00:00:24] So don't wait, now get in and activate the My Magenta app on Telekom.
[00:00:33] Welcome to Tech News Briefing, it's Friday August 30th.
[00:00:37] I'm Zoe Thomas for the Wall Street Journal.
[00:00:40] Legions of tech-savvy Gen Zers can do anything on a computer, except type without looking at the keys.
[00:00:48] How did young people who spend so much time online fail to learn typing?
[00:00:53] Plus, car makers are grappling with the ethical dilemmas self-driving vehicles will face on the road.
[00:00:59] We'll tell you how companies, bioethicists and engineers are approaching some of these thorny questions.
[00:01:11] But first, over the past 25 years, the number of U.S. high schools teaching typing has fallen drastically.
[00:01:20] According to the U.S. Department of Education, in the year 2000, 44% of students who graduated high school
[00:01:26] took a keyboarding course.
[00:01:28] By 2019, just 2.5% of high school graduates had.
[00:01:33] That means that the generation known for being digital natives could be typing like someone who was just introduced to a computer.
[00:01:41] Here to tell us what's going on is our reporter, Georgia Wells.
[00:01:44] Georgia, why did the number of high school students taking a keyboarding course drop by so much?
[00:01:50] Educators tell me that schools have gotten rid of courses and topics that are not explicitly tested on standardized exams.
[00:01:58] And so that typing is not the only casualty there, but typing is one of the significant ones.
[00:02:04] People assume that Gen Z kids, because they're these digital natives, will know how to type.
[00:02:10] But clearly, kids don't just wake up one morning knowing how to type.
[00:02:13] So that's kind of the perniciousness of these assumptions.
[00:02:16] What have some schools done to try and reverse this trend?
[00:02:20] So in Oklahoma City, this woman, Christine Miller, she noticed that in third grade,
[00:02:25] the kids didn't know how to type, but they were given these laptops and teachers were like,
[00:02:29] oh my goodness, I cannot fit in a typing course on top of the curriculum.
[00:02:33] And so she implemented this spelling bee style competition called the Keebee,
[00:02:38] where she hoped that students would start practicing typing on their own accord to try and perform well in this test.
[00:02:44] And so on the day of the test, Rumble the Bison, who's the mascot for the Oklahoma City Thunder,
[00:02:50] showed up to cheer on the students.
[00:02:52] The students got there, and the fastest student clocked in at 91 words a minute, which is really quite fast.
[00:02:59] So what are some of the ways students are typing their work if they're not using a traditional keyboard?
[00:03:05] Yeah, so a lot of school districts give out iPads.
[00:03:07] So there was a student I interviewed who described becoming so proficient on typing on the screen of an iPad
[00:03:13] that she actually preferred typing on an iPad to a laptop.
[00:03:17] And so this showed me that a lot of students were getting their work done,
[00:03:20] even if it didn't mean they were sitting at a laptop typing with traditional home row form.
[00:03:26] Do we know how much work is being submitted that's done on mobile devices?
[00:03:31] So there's this company in structure, and they make the software called Canvas.
[00:03:36] And Canvas is this tool that students use to submit work, or schools use to assign students work.
[00:03:41] Canvas noted that between March and May of this year,
[00:03:44] that 39% of the assignments that students were submitting, they were submitting from mobile devices.
[00:03:50] And what was really interesting was they noticed that students and teachers are logging onto Canvas on different devices.
[00:03:58] 90% of teachers were completing their work on a computer,
[00:04:03] even if 39% of students are uploading their stuff from mobile.
[00:04:07] And so they noted that there's this disconnect between how students and how teachers are approaching the same work.
[00:04:13] So what's the case for learning how to touch type?
[00:04:16] Some of the teachers I spoke with said that one of their fears is that a lot of young people use the devices in their lives as consumption tools,
[00:04:25] rather than tools to create.
[00:04:27] And so they were arguing that if students learned to type faster and sooner
[00:04:30] that they'd be more comfortable creating content on their devices,
[00:04:34] rather than just having a sit back experience.
[00:04:37] That was our reporter, Georgia Wells.
[00:04:39] Coming up, as car makers work to get more self-driving vehicles on the road,
[00:04:43] companies are grappling with how to train their AI to make life and death decisions.
[00:04:49] That's after the break.
[00:05:06] Many car makers say autonomous vehicles will make driving safer.
[00:05:11] The artificial intelligence powering self-driving cars doesn't get tired, distracted or intoxicated,
[00:05:17] factors that can lead to crashes.
[00:05:19] But that doesn't mean the companies and people developing self-driving cars won't have to grapple with ethical questions.
[00:05:26] With more on that, we're joined by Brett Burke who wrote about this for the Wall Street Journal.
[00:05:31] So Brett, it's important that autonomous vehicles be able to quote-unquote see pedestrians to avoid hitting them,
[00:05:38] but there are questions about bias being introduced here.
[00:05:41] There have been shown to be biases in the training data that some of these artificial intelligence systems have been trained on.
[00:05:48] And this is especially problematic in the case of autonomous vehicles
[00:05:53] because the training set oftentimes does not include a deep reservoir of people of color.
[00:05:59] So the camera systems can sometimes be limited in some ways in their ability to see people with darker skin.
[00:06:06] What are some of the possible solutions for this?
[00:06:09] Better cameras and better training data.
[00:06:11] This is one of the issues that we see in general with AI,
[00:06:14] is that it's only as good as the data that it's trained on.
[00:06:17] What are some of the other ways that a self-driving car might detect a human or something else nearby?
[00:06:23] There are a number of different systems that these vehicles will utilize.
[00:06:27] There's vehicle to vehicle communications, what we call V2V,
[00:06:30] that allows them to communicate with each other so they can inform each other about their planned path,
[00:06:36] what they're going to do, what they're seeing.
[00:06:37] There's vehicle to infrastructure, V2I,
[00:06:40] which allows them to communicate with traffic lights or sometimes with cameras, traffic cameras,
[00:06:45] or other elements of infrastructure, draw bridges, things like that so they know what's coming up.
[00:06:50] And then there's work on this idea of V2P, which is vehicle to pedestrian communications,
[00:06:55] which would allow autonomous vehicles and the AI to run them to detect and ensure
[00:07:00] that what they're seeing is an actual person as opposed to maybe a mannequin or a humanoid delivery robot.
[00:07:06] How are companies approaching this?
[00:07:08] For example, Waymo, which is Alphabet's self-driving subsidiary,
[00:07:11] what has it said about dealing with animals potentially on the road?
[00:07:16] Well, for large animals, its goal is to reduce injury-causing potential for humans on board.
[00:07:21] It would take avoidance maneuvers, if necessary, to try to stop any impact from happening with a large animal.
[00:07:28] With smaller animals, when we talked to Waymo, they said that they recognize that evasive action
[00:07:33] for certain categories of roadside animals or debris or whatever can be at risk.
[00:07:40] So a lot of it is about analyzing risk and trying to minimize risk for the people on board,
[00:07:45] people that are in surrounding the cars and the vehicle itself.
[00:07:50] What do bioethicists make of this?
[00:07:53] For this story we talked to Peter Singer who's a famed bioethicist from Princeton.
[00:07:57] His general notion is that some respect for animals is better than no respect for animals.
[00:08:02] So if there are decisions to be made that privilege house pets or farm animals or whatever it might be,
[00:08:09] he finds that to be a better moral position than just saying disregard any animal.
[00:08:14] You know, when we're talking about taxis, people often want to get to their destination as quickly as possible
[00:08:20] and safely as possible.
[00:08:23] And then companies want to move as fast to pick up as many passengers as they can,
[00:08:27] but they also want to avoid accidents that can slow things down.
[00:08:31] You spoke with the CTO of Mobileye, that's a unit of Intel,
[00:08:35] that deploys autonomous vehicle software and hardware.
[00:08:38] How is that company approaching this issue?
[00:08:41] The folks that I talked to from Mobileye were sort of talking about this idea of a trade-off between safety and expedience.
[00:08:49] So, you know, if you're as safe as you possibly can be,
[00:08:52] that could lead to obviously fewer accidents but also traffic snarls
[00:08:57] and, you know, if you maximize speed, you'll make more trips.
[00:09:01] People will get to their destinations more quickly,
[00:09:03] but then there's the risk of more incidents if things are moving too quickly.
[00:09:08] I mean, is there a way that this might be approached with, say, different kinds of cars?
[00:09:12] Well, if you think about some of the things that differentiate vehicles for individuals
[00:09:17] when they're thinking about an automotive purchase,
[00:09:19] an autonomous vehicle and the AI that runs it could be trained to deliver a driving experience
[00:09:24] that differs from automotive brand to automotive brand.
[00:09:28] So, an autonomous sports car might drive a little bit more aggressively
[00:09:31] or choose a route that shows off its handling capabilities.
[00:09:36] Similarly, maybe an off-road vehicle or something like that might be programmed to take a more scenic route.
[00:09:42] One of the big questions seems to be, should autonomous vehicles be held to a higher standard than human drivers?
[00:09:49] How are companies approaching this?
[00:09:51] When a human has an incident, right, we can sort of say,
[00:09:53] well, they're only human. It was an accident.
[00:09:56] We're going to hold autonomous vehicles to a higher standard
[00:09:59] and we're going to be less forgiving if they're being run and programmed by a company where these decisions are being made.
[00:10:05] This isn't just like incidental things that are happening.
[00:10:07] There's programming that needs to happen in order for these decisions to be made the way that they are.
[00:10:11] And we can see that with Tesla a couple of years ago.
[00:10:14] They got dinged by the National Highway Traffic Safety Administration
[00:10:17] for their quote-unquote full self-driving service, which allowed vehicles to roll through stop signs.
[00:10:24] Obviously a lot of humans behave this way, but when it's programmed into an autonomous vehicle,
[00:10:30] it becomes a different issue.
[00:10:32] And so is this a time now with the coming of this to really be considering how we can think about our roads in a way that sort of put people first?
[00:10:41] Alright, that's Brett Burke. Thanks for joining us for this conversation.
[00:10:44] Thanks so much again for having me. I appreciate it.
[00:10:48] And that's it for Tech News Briefing.
[00:10:50] Today's show was produced by Zoe Colkin.
[00:10:52] On your host Zoe Thomas, Jessica Fenton and Michael Lavalle wrote our theme music.
[00:10:57] Our supervising producer is Catherine Millsop.
[00:10:59] Our development producer is Aisha Al-Muslim.
[00:11:02] Scott Salloway and Chris Sinsley are the deputy editors.
[00:11:05] Anphelana Patterson is the Wall Street Journal's head of news audio.
[00:11:08] We'll be back this afternoon with TNB Tech Minute as usual.
[00:11:12] But then we're off Monday for Labor Day in the U.S.
[00:11:15] Thanks for listening.