Can we say developing AV technologies is a disruptive innovation? Why

The picture in autonomous vehicle (AV) or self-driving vehicle sector The long, winding road for driverless cars Carmakers like to talk about autonomous vehicles as if they will be in showrooms in three or four years' time. The rosy picture they paint suggests people will soon be whisked from place to place by road-going robots, with little input from those on board. Autonomous vehicles will end the drudgery of driving, people are told. With their lightning reactions, tireless attention to traffic, better all-round vision and respect for the law, autonomous vehicles will be safer drivers than most motorists. They won't get tired, drunk, have fits of road rage, or become distracted by texting, chatting, eating or fiddling with the entertainment system. The family autonomous vehicle will ferry children to school; adults to work, malls, movies, bars and restaurants; the elderly to the doctor's office and back. For some, car ownership will be a thing of the past, as the cost of ride-hailing services like Uber and Lyft tumbles once human drivers are no longer needed. Going driverless could cut hailing costs by as much as 80%, say optimists. Welcome to the brave new world of mobility-on-demand. All these things may come to pass one day. But they are unlikely to do so anytime soon, despite the enthusiasm of people such as Elon Musk, the boss of Tesla, a maker of electric cars. Within two years, he says, people will be napping as driverless vehicles pilot them to their destinations. Mr Musk has defied conventional wisdom before, and proved critics and naysayers wrong. In this case, however, too many obstacles lie ahead that are not amenable to brute-force engineering. It could be a decade or two before Autonomous vehicles can transport people anywhere, at any time, in any conditionsand do so more reliably and safely than human drivers. Consider how long it has taken for something as simple as electrically powered cars to carve a niche for themselves. After a couple of decades, hybrid and pure-battery vehicles still account for no more than 2% of new-car sales in most countries. Battery prices and storage capacities are now approaching a point where sales could feasibly take off. But even using the most optimistic of assumptions (say, electrics accounting for half of new-car sales), it would be 2035 at the earliest before they represented half the vehicles on American roads. Expect fully autonomous vehicles to face an equally long and winding road. To put matters in perspective, most cars on the road today require the driver to do practically everythingsignaling, steering, accelerating, braking, watching the traffic ahead, to the sides and to the rear. This is Level 0 motoring on the scale of autonomous vehicles devised by the Society of Automotive Engineers (SAE) in America. Vehicles equipped with rudimentary forms of driver-assistance, such as cruise control or reversing sensors, are classed as Level 1. Fitted with wide-angle cameras, GPS sensors and short-range radars, Level 2 vehicles can adapt their speed to the surrounding traffic automatically, maintain a safe distance from the vehicle ahead, keep within their own lane, and even park themselves occasionally. For short stretches of time, the driver's hands may be removed from the steering wheel and his feet from the pedals. But the driver must be ready to take full control of the vehicle at any instant. Tesla's Autopilot system is classed as Level 2 technologyor was until it was rolled back recently to Level 1 for safety reasons. In the accident that killed a Tesla driver in Florida in 2016, that driver either failed to respond in time to avert disaster, or mistakenly assumed that Autopilot meant more (as its name implied) than mere driver-assistance. Tesla continues to include the Autopilot sensors and software in its cars, but has deactivated the system while further testing is undertaken. The company plans to re-activate it in 2020 or thereafter. 2 Level 3 autonomous driving is even more controversial. The main difference is that, while the driver must still remain vigilant and ready to intervene in an emergency, responsibility for all the critical safety functions is shifted to the car. This has a lot of engineers worried. Experience has not been good with control systems that relegate the operator to a managerial role whose only job is to intercede in the case of an emergency. It was this sort of thinking that allowed an accident at a nuclear power plant at Three Mile Island, in 1979, to escalate into a full-blown meltdown. Plant operators failed to react correctly when a valve stuck open and caused the reactor to lose cooling water. They then made matters worse by overriding the automatic emergency cooling system, thinking there was too much water in the reactor rather than too little. The accident report blamed inadequate operator training and a poorly designed computer interface. Similar human failings have led to countless airline accidentsmost recently, the Asiana Airlines crash at San Francisco in 2013. Over-reliance on automation and lack of understanding by the pilots about when they needed to intervene were cited as important factors contributing to the Asiana crash. Some carmakers fear thateven more than reactor operators or professional pilotsuntrained motorists may only worsen the problem when suddenly required to take control of an otherwise fully automated system. Ford believes it is better to skip Level 3 altogether, and go straight to Level 4, even if it takes longer. In theory, Level 4 technology should be safer. Such vehicles will carry out all critical driving functions by themselves, from the start of the journey to the end. The only proviso is that they will be restricted to roads that they have been designed for. This means routes which have been mapped in three dimensions and "geofenced" by GPS signals, to prevent Autonomous vehicles on them straying outside their designated zones. Ride-sharing services like Lyft and Uber are likely to be the first to operate Level 4 vehicles. In fully autonomous Level 5 motoring, the vehicles have to perform in all respects at least as well as human driversin short, they must be capable of going anywhere, in every conceivable condition, and be able to cope with the most unpredictable of situations. That means travelling on dirt tracks off the map, in blizzards, thunderstorms or pitch darkness, with animals bursting out of bushes, children chasing runaway balls and crazy people doing crazy things. To fulfill their promised role, Level 5 cars and lorries will have to do all this and more. The most crucial piece of technology needed to make that happen is lidar (light detection and ranging). Lidar uses pulses of laser light flashed from a rotating mirror on a vehicle's roof to scan the surroundings for potential obstacles. Unlike video cameras, lidar cannot be dazzled by bright light. Nor can it be blinded by darkness. It is far more accurate than radar at measuring the distance and speed of objects. Better still, it provides an image in three dimensions. Clever mathematics can allow lidar sensors to tell whether an object is hard or softin other words, whether it is another vehicle or a wayward pedestrian. The Autopilot in Tesla's cars does not rely on lidar, though it may yet have to if it is to match the resolution of other ranging and detection systems. While not cheap, lidar sensors are available from a number of suppliers. But as far as autonomous vehicles go, the real value lies not in the laser hardware, but in the software that combines lidar images with signals from radar detectors and video cameras, and overlays the resulting 3D map with GPS data. Waymo, an autonomous-vehicle firm set up by Alphabet (Google's parent), is believed to have the most sophisticated lidar systems of all. They are expected to be the jewel in the crown when Waymo comes to license its autonomous technology to carmakers around the world. 3 Hence a court battle that has raged in San Francisco, with Waymo accusing Uber of having stolen its intellectual property and copied its lidar designs. In a ruling on May 15th, 2017, Uber was ordered to return the stolen property and for the alleged culprit to cease work on the disputed hardware. There is now a real possibility of criminal charges being filed. That would seriously hamper Uber's plans to field a fleet of ride-sharing cars without costly drivers. Before this can happen, however, numerous matters need resolving. Local governments will have to spend money making roads more (autonomous vehicle) AV-friendly. Besides that, who will be responsible for all the vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) low-latency wireless networks needed to manage the driverless traffic? On another issue, the courts have yet to determine how Autonomous vehicles should share the road with unpredictable human drivers. Who will be held liable when the inevitable accidents happen? The AV owner? The manufacturer? The software supplier? And how will Autonomous vehicles be shielded from cyber attack? Meanwhile, the legal and ethical discussions around letting algorithms resolve human dilemmas concerning life and death have barely begun. The enthusiasts are right on one thing, though: testing is paramount. Waymo has logged more than 2m miles (3m km) of autonomous driving. But that is nowhere near enough to gauge whether autonomous vehicles are safe enough to let loose on the public. According to statistics from America's Bureau of Transportation, there were about 35,000 fatalities and over 2.4m injuries on American roads in 2015. That may sound a lot but, given that Americans drive three trillion miles a year, accident rates are remarkably low:1.12 deaths and 76 injuries per 100m miles. Because accidents are so rare (compared with miles travelled), autonomous vehicles "would have to drive hundreds of millions of miles, and sometimes hundreds of billions of miles, to demonstrate their reliability in terms of fatalities and injuries," says Nidhi Kalra of RAND Corporation, a think tank in California. At present, there is no practical means for testing the safety of Autonomous vehicles before their use becomes widespread. For many, that is a scary thought. Nor is there any consensus on how safe Autonomous vehicles should be. "Safer than human drivers" ought to be a minimum requirement. Some would go further and require roadgoing robots to present no threat whatsoever to human life. That would imply it is acceptable for humans to make mistakes, but not for machines to do so. Such safety issues will have to be resolved before any regulatory framework for autonomous vehicles can be put in place. All of which raises questions for consumers to ponder. The most obvious one is economic. According to Fitch, a financial-ratings agency, the average car spends 96% of its usable life parked in a garage or on the street. When maintenance, depreciation, insurance and running costs are totted up, cars are the most underused asset consumers own. So, what happens when people have an alternative that is cheaper than owning a car but just as convenient? Clearly, some will opt to hail driverless vehicles instead of having a car. Carmakers could thus find themselves selling fewer vehicles to individuals, and more to operators of driverless fleets, who will run them 24 hours a day, seven days a week, and scrap them after a year or two. The coming era of "transport-as-a-service" suggests motor manufacturers will need to rethink the way they do business. So as not to be blind-sided, General Motors has invested $500m in Lyft and forked out nearly $1 billion to acquire Cruise Automation, an autonomous-vehicle developer in Silicon Valley. Meanwhile, Ford has replaced its automotive-industry stalwart of a boss with an outsider brought in last year to oversee experiments with self-driving cars and ride-sharing services. FiatChrysler has been using 100 minivans to test autonomous technology supplied by Waymo. 4 Most carmakers have plans to start testing the market with Level 3 or possibly Level 4 autonomous vehicles around 2021. Such Autonomous vehicles may still have steering wheels and pedals, and be able to drive autonomously only on designated roads. The majority are likely to be bought by ride-hailing services. Consumers wanting the flexibility and freedom of full Level 5 vehicles will have to wait a good deal longer. Come that day, though, the choice will be drive or be driven. The betting is that a surprising number of people will still want to drive themselveswhether out of mistrust of the machine or the satisfaction that comes from having total control. Who's self-driving your car? With its successful test of robo-taxis on the streets of Pittsburgh in September 2016, Uber has dominated recent headlines on autonomous vehicles. But behind the scenes three groups technology giants such as Uber, carmakers and a whole fleet of autoparts suppliersare in a tight race. Each is vying to develop the hardware and software that make up the complex guts of a self-driving vehicle. A couple of years ago tech firms appeared well ahead in this battle. But, Uber aside, they have dabbed the brakes of late. The recent departure from Google of Chris Urmson, the company's figurehead for autonomous vehicles and the man who once promised it would put self-driving cars on the road in 2017, is a significant reversal. The recent slimming of the team at Apple that is devoted to building an autonomous electric car, also shows that tech firms are not having it all their own way (though Apple's possible tie-up with McLaren, a British maker of sports cars and Formula 1 racing team, would be one way to put its carmaking ambitions back on track). Carmakers, meanwhile, are making more of the running after a slow start. Despite recent safety concerns, Tesla, an electric-car maker, is making progress with its Autopilot system. In 2017 Volvo, which is also working with Uber to get cars to drive themselves, will test selfdriving cars by handing them for the first time to a select group of ordinary motorists. And in August 2017, Ford said it would launch a fully-autonomous car, without steering wheel or pedals, for car-sharing schemes by 2021. All parties recognize that the biggest profits from autonomy will come from producing an "operating system"something that integrates the software and algorithms that process and interpret information from sensors and maps and the mechanical parts of the car. Tech firms probably have the edge here. But carmakers and suppliers are not giving up easily. So they are involved in a bout of frenzied activity to keep control of the innards of self-driving cars. In July, for example, BMW, Mobileye, an Israeli supplier that specializes in driverless tech, and Intel, the world's biggest chipmaker, said they were joining forces. Another strategy for carmakers is to develop autonomous driving in-house. They are hoovering up smaller firms that have useful self-driving technology, notes Andrew Bergbaum of AlixPartners, a consulting firm. Ford has put money into a lidar company, and into another that sells mapping services. It has also acquired two other firms that specialize in machine-learning and other artificial-intelligence technology. The losers in this race look likely to be the big parts-makers, whose relationship with their main customers could become strained. Over time carmakers have largely ceded to them the job of developing new technology. If they turn back the clock and reintegrate vertically that may leave less business for the suppliers. The tech giants still have huge advantages. As well as their financial resources, they are in the best spot to claim the big profits from the operating system. Apple's plans to build a car may 5 be swiftly revived if it buys McLaren. And Google is ahead in machine-learning, the vital element in developing algorithms that will eventually replace drivers. But carmakers are coming up surprisingly fast on the inside lane. What's going on in the self-driving car sector The whizzy gadgets for geeks to goggle at during CES, an annual consumer-electronics show in Las Vegas, have typically been small enough to pick up. But they have been joined in recent years by an increasing number of cars. The Detroit motor show, America's biggest and glitziest, starts later in January 2016, but many in the car industry now regard CES, which opened in the same month, as a more important event. Mary Barra, GM's boss, unveiled a new production version of its Bolt electric car at Las Vegas in February 2016. Incumbent manufacturers are recognizing the double threat posed by technology, as carsharing takes off and driverless vehicles come closer. First, some people who might hitherto have wanted to own a car may no longer do so, cancelling out the growth the motor industry might otherwise have expected from the rising middle classes in developing countries. Second, technology firms may be better placed than carmakers to develop and profit from the software that will underpin both automated driving and vehicle-sharing. Some of these firms may even manufacture cars of their own. In a report ahead of the Las Vegas and Detroit shows, Morgan Stanley, an investment bank, said the motor industry was being disrupted "far sooner, faster and more powerfully than one might expect." It predicted that conventional carmakers would scramble in the coming year to reinvent themselves. As if to demonstrate this, shortly before CES opened, GM announced a $500m investment in Lyft, a ride-sharing service. A rumored tie-up between Ford and Google to produce driverless cars failed to materialize at the show, but even the rumors underlined the disruption that tech firms are bringing to the motor industry. And other partnerships were announced: Ford is teaming up with Amazon to connect its cars to sensor-laden smart homes. It was also revealed at CES that Toyota would adopt Ford's in-car technology, which is a competitor to Apple's CarPlay and Google's Android Auto, to access smartphone apps and other features. That is not the only example of carmakers joining forces to avoid being beholden to the tech giants. In August 2016, BMW, Daimler and Volkswagen's Audi division jointly bought Here, a mapping service, from Nokia, to ensure that carmakers have an independent provider rather than having to depend on Google Maps. Nevertheless, carmakers are also teaming up with tech firms because each has something the other needs. Building and marketing cars, and dealing with safety and emissions regulators, is tricky. Tech firms could copy Tesla, which has built its own electric cars for more than a decade. Apple, which is said to be planning an electric car, may try to have them made in the same way as it does its iPhones, outsourcing to a contract manufacturer. But a more obvious route is to ally with an established carmaker. Carmakers also have lots to learn. Most are working on making their vehicles either fully or partly self-piloting, and a number are running their own car-sharing experiments. But Google remains the leading exponent of autonomous driving. Its robotics, drones and search engine all contribute expertise that helps to guide a driverless car down the road avoiding pedestrians, obstacles and other vehicles, using computing power and sophisticated software to interpret masses of data received both from the car's on-board sensors and from external sources through wireless connections. Yet if the tech firms have much to gain as they muscle in to the motor business, the carmakers are wary of what they have to lose. Profits may seep away towards the producers of 6 the software and the owners of the data, and away from the makers of the hardware. Hitherto, new carseven quite modest oneshave tended to be bought as status symbols and expressions of personal style, but if consumers become more interested in what software and entertainment systems a car can run, rather than what it looks like, the industry's whole business model may come apart. Ride-sharing, car clubs and other alternatives to ownership are already growing fast. Young city-dwellers are turning their backs on owning a costly asset that sits largely unused and loses value the moment it is first driven. Carmakers insist that such consumers are merely deferring buying a vehicle, pointing to the fact that people continue to drive at an older age than they used to. But the pronouncements of motor-industry bosses suggest that doubts are creeping in. At CES Mark Fields, Ford's CEO, said that it would in future be "both a product and mobility company". Membership of car clubs, which let people book by app for periods as short as 15 minutes, is growing by over 30% a year, according to Alix Partners, a consulting firm, and should hit 26m members worldwide by 2020. Competition is intense. ZipCar, owned by Avis Budget, a conventional car-hire firm, is thriving. More carmakers are copying Daimler's Car2Go and BMW's Drive Now apps. Earlier this year Ford began testing both a car-sharing service in America and a car club in Britain. Daimler reckons its scheme is profitable. But such services are unlikely ever to match the returns, especially for premium makers, from selling vehicles. At the same time, app-based taxi services such as Uber and its Chinese counterpart Didi Dache, which are often cheaper and more efficient than conventional cabs, are also growing quickly. Once these are able to dispense with drivers for their vehicles, the taxi, car-club and carsharing businesses will in effect merge into one big, convenient and affordable alternative to owning a car. So when will the fully autonomous car hit the showrooms? Google, whose cars have done 1.3m test miles (2.1m km) on public roads, once promised 2018, whereas most analysts reckoned the 2030s more plausible as carmakers introduced automated-driving features in stages. Now, Mr Fields is talking about autonomous cars being ready to roll by 2020. More conservative car bosses add five years. Barclays, another bank, forecasts that the fully driverless vehicle will result in the average American household cutting its car ownership from 2.1 vehicles now to 1.2 by 2040. A selfpiloting car may drop off a family's breadwinner at work, then scuttle back to pick up the kids and take them to school. The 11m or so annual sales of mass-market cars for personal ownership in America may be replaced by 3.8m sales of self-driving cars, either personally owned or part of taxi fleets, Barclays thinks. Driverless cars still have problems in bad weather. They may struggle to recognize that light shining off a puddle is harmless or guess that a pedestrian is about to step into the traffic without looking. But sophisticated systems for hands-free driving on motorways, and for automated parking, are already available on a number of manufacturers' models. Fully driverless cars will ferry workers round GM's technical center in Detroit in late 2016. Convincing regulators to allow fully driverless cars onto the streets is the next hurdle. Insurers and consumers also need to be won round. If self-driving cars can be introduced first on private roads or designated areas of cities to prove their worth in avoiding accidents and reducing congestion, that might help. Within the industry, the big question is not whether this future will arrive, but whether tech firms or carmakers will grab the spoils. Will the sign on the dashboard say Ford (powered by Google) or Google (powered by Ford)? Collaborations among multiple players 7 Just over a year since BMW Group teamed up with chipmaker Intel and sensor firm Mobileye to develop a driverless car platform in 2016, the alliance has signed up another major car manufacturer: Fiat Chrysler. The addition of Fiat Chrysler will give the alliance a big potential boost in the North American market. "In order to advance autonomous driving technology, it is vital to form partnerships among automakers, technology providers and suppliers," Fiat Chrysler CEO Sergio Marchionne said in a joint statement. Since the alliance's formation in mid-2016, BMW and Chinese tech giant Baidu called off their partnership on driverless car research, and Intel bought Mobileye, an Israeli company whose technology enables "computer vision," for $15.3 billion. Intel's deal aimed to address a perceived deficit in autonomous driving technology, an area that is likely to be one of the biggest growth for chipmakers for years to come. A fully autonomous car is likely to need up to $500 worth of semiconductors, according to some estimates. So far, the companies involved in the alliance have announced two tranches of self-driving car tests that are scheduled to take place by the end of 2017: a 40-car test that was revealed at the start of the year, and a 100-car test that Intel just announced. "The future of transportation relies on auto and tech industry leaders working together to develop a scalable architecture that automakers around the globe can adopt and customize," said Intel chief Brian Krzanich. The platform that the companies are developing is intended for use by multiple car manufacturers, and indeed Wednesday's statement reiterated their invitation for more automakers-and tech firms-to join the party. The system should be able to support various degrees of autonomy, ranging from what the industry calls "Level 3" (which still requires a human driver for many tasks) all the way up to "Level 5" (which provides a fully hands-off experience in all conditions).Other notable partnerships in the autonomous car field include those between Daimler and Uber, and General Motors and Lyft. Actions taken by Uber and Lyft. Both Lyft and Uber have invested in autonomous vehicles development. But until recently, Uber had a significant head start in building its own technology, partnering with others and launching self-driving pilots. More recently, though, Lyft has started to catch up. The company inked its first driverless-car partnership, with General Motors, in early 2016, and more recently added another deal, with MIT spinoff NuTonomy. In May 2017, it came out that Lyft had another partner none other than Waymo, the litigious thorn in Uber's side. "It was an easy and straightforward partnership to make," says John Krafcik, CEO of Waymo. And later in July 2017, Lyft announced its biggest leap yet toward an autonomous future the development of a suite of technologies, both hardware and software, that will allow any manufacturer to turn its vehicles into driverless machines and to easily integrate with Lyft's network of passengers. With the new tech, Lyft is offering a new pitch to potential partners: We're in it to win it, and we'll give you tools that'll make it easier for you to win too. Uber is quick to note that self-driving is core to its mission (it has been developing many of the above technologies for a couple of years now) as well, and that it is still growing at a morethan-healthy clip despite the recent controversies. "Our business is stronger than ever, and we're keeping our heads down to build the best and most innovative products out there," a spokesperson said. Leapfrogging Uber won't be easy. But driverless tech could elevate Lyft to parity, or betternot to mention make it profitable. (Lyft reportedly brought in revenue of about $700 8 million and lost about $600 million in 2016; Uber says it lost $2.8 billion on revenue of about $6.5 billion.) Is there room for two giants to coexist in a driverless future? Perhaps. But in this high-pressure race, the winner gets far more than bragging rights. Lyft is also announcing the launch of a self-driving division called Level 5. Based in Palo Alto, the Level 5 facility will house hundreds of engineers that will work on developing the open self-driving system. Lyft's efforts in the self-driving space will help it take on the market's 800- pound gorilla. As Lev-Ram writes, leapfrogging Uber won't be easy but "driverless tech could elevate Lyft to parity or better." GM made the first move GM reveals barn-sized truck at Detroit motor show. What else is new, you might now ask. But the launch on January 20, 2018 of the Chevrolet Silverado, a pickup that will go on sale at the end of the year, highlights a surprising turnaround for America's largest carmaker. The good news is not just the Silverado's outsized margins, which are important for a firm that relies heavily on trucksafter Mary Barra, GM's boss, gave an ebullient performance at an investors' conference that coincided with the motor show, the release of GM's quarterly results on February 6th are likely to include record profits. It is also that the money thrown off by vehicles such as the Silverado will help the firm navigate the tricky terrain that lies ahead of all the world's big carmakers. One task is to ensure that their current business of selling vehicles with internalcombustion engines stays healthy. At the same time, they must prepare for a future of electric and autonomous cars (EVs and AVs), which threaten to up-end business models that have endured for a century. Not so long ago, GM and its peers seemed to be on a path to extinction. Technology firms such as Alphabet, Uber and other pushy newcomers had started a race to develop software that would control driverless cars and to offer ride-hailing and ride-sharing services that are expected to thrive at the expense of car ownership. In April 2017 GM's market value was overhauled by Tesla's, a firm that makes just tens of thousands of flashy EVs a year, compared with the millions of vehicles rolling off GM's production lines. Sentiment has changed dramatically. Since April GM's share price has surged by 28%, giving the firm back the lead. By contrast, Tesla has struggled with the nuts-and-bolts of carmaking. Production-line problems have hampered a big roll out of its mass-market Model 3. Analysts at Barclays, a bank, say that GM is more "evolving mammal than...dying dinosaur". One reason for the reversal of fortunes is that GM has convinced investors that its current business is in fine fettle. The cash generated by the Silverado and a range of new pickups will help pay for big investments in EVs and AVs. It does not hurt GM's case that Ford, its main rival in Detroit, is struggling. Jim Hackett, a new boss brought in because of his technology knowhow, oversaw a lacklustre relaunch in October that was sketchy on Ford's vision for the future of transport services. On January 24, 2018 the firm reported disappointing quarterly results, dashing hopes for quick improvement. In contrast, GM is already well on the way to reshaping itself. For starters, it has diverted resources to where it is a market leader. It has got rid of unprofitable businesses around the world, a process that culminated in a decision last March to sell Opel, its loss-making European carmaker, to France's PSA. At the same time, GM has invested heavily in new pickups, such as the Silverado. Cadillac, GM's premium brand, may look like an exception to this happy rule. Sales of just 350,000 cars in 2017 puts it far behind its German rivals. Yet sales have doubled since 2010 9 and it has grown faster than any of them in recent years. Although the firm does not disclose the information, analysts at Morgan Stanley reckon that Cadillac could be worth $13bn, around 20% of GM's current value. Johan de Nysschen, Cadillac's boss, admits he runs a "challenger brand", but sniffs an opportunity. The upheaval created as carmakers grapple with new business models means that "everyone has to start again". The most important reason for GM's comeback, though, is its success in convincing investors that it is a leader not just among established carmakers, but among tech firms, too. It has rapidly accelerated from the position of an also-ran in the field of autonomous vehicles to apparent leader. A scorecard issued annually by Navigant, a consultancy, puts GM ahead of the AV pack of carmakers and tech firms, with Alphabet's Waymo in second place. That GM is ahead of Silicon Valley's risk-takers may seem surprising. But earlier investments, which were once looked on with scepticism, seem to be paying off. Alan Batey, GM's president for North America, points to the manufacturing of mass-market long-range EVs, where the firm has a lead. The Chevy Bolt, the world's first such vehicle, has been on sale for over a year, beating Tesla's Model 3 and the new Nissan LEAF to market. The Bolt is supposed to be the basis for an ambitious autonomous ride-sharing business. On January 12th GM announced the latest version of its Cruise AV, a Bolt-based robotaxi without a steering wheel or pedals. GM plans to use it to launch a commercial scheme in several cities, starting next year. Rival tech firms and carmakers are only running, or are planning to launch, small test projects. Revenge of the robotaxis When GM paid $1bn in 2016 for Cruise, an artificial-intelligence startup, many analysts wondered whether it was throwing away money. But the marriage of cutting-edge technology and large-scale manufacturing seems to be paying off. The carmaker has learned to be more nimble; Cruise has picked up how to make its fiddly technology robust enough for the open road. As a result, GM can now mass-produce self-driving cars, says Dan Ammann, second-incommand to Ms Barra. Scale will help steeply to reduce the cost of sensors, which are the key components of an AV. The firm is being rewarded because, unlike other carmakers, it has assembled all the parts of the puzzle you need to build new transport services, says Stephanie Brinley of IHS Markit, a consultancy. But even if GM is no longer a dinosaur, risks remain. In particular, it may be too bullish in its estimate of the market for robotaxis and it may be placing too much faith in the benefits of being the first to market. The company expects demand to expand quickly. Costs of ride-hailing services, it predicts, will fall from $2.50 a mile now to about $1 as the main expensethe driveris eliminated. In America alone it would be able to tap a market worth around $1.6trn a year (representing three-quarters of all miles travelled) as drivers are lured from their cars to robotaxis. But what Mr Ammann calls this "very big business opportunity" comes with an inconvenient corollary. As car buyers become car users, GM's legacy business supplying vehicles to drive will decline accordingly. Critics think that GM may have accelerated too swiftly and that it will have to endure years of losses before robotaxis take off. Even if things move fast, points out Berenberg, another bank, GM may not be the one to benefit. The main constraint in growing a ride-hailing business now is acquiring drivers. But when these are eliminated, capital will be the only limit. And that could mean huge fleets of robotaxis chasing passengers, forcing prices down. Riders may then 10 choose a brand they recognize, such as Uber and Lyft, rather than Maven, GM's ride-hailing business. If so, being first would confer little advantage. And yet, if carmakers do not want to accept their fate passively, they have little choice but to remodel themselves. The outsized Silverado and the sensor-packed Cruise AV show that GM has the present in handand that it is at least doing its best to safeguard its future. Some thoughts on self-driving cars The tech superpowers are teaming up for a good reason: Autonomous vehicles are our future. While significant technical challenges remain unsolved, Autonomous vehicle technology is improving rapidly. Soon technological capability won't be the greatest impediment to adoption; societal friction will be. This friction will delay full autonomy for at least a decade, or however long it takes for the tech community (which hasn't always been particularly empathetic) to collaborate with policymakers, regulators, insurance providers, and consumer advocates to address the significant social, regulatory, and legal challenges Autonomous vehicles will create. Autonomous vehicles can dramatically reduce vehicular injuries and deaths, but we don't yet know how people will react to the inevitable mistakes that will occur. Tesla's self-driving technology was at first blamed and then cleared for a fatal crash in Ohio, but Uber's self-driving cars missed at least six red lights in San Francisco last year. Even though society accepts car accidents as an unavoidable (and horrible) risk associated with mobility, machine error will be judged much more harshly than human error due to the expectation of precision. With artificial intelligence (AI) playing an increasing role in autonomous vehicle technology, self-driving vehicles will inevitably make "wrong" choices when presented with options of who will live and who will die. Many companies and institutions are tackling the ethical side of decisions made by self-driving cars, such as MIT's Moral Machine (a judgment crowdsourcing platform) and Crowdflower's approach to machine learning with human-in-theloop AI. Over time, more segments of society will need to contribute to these morally and legally challenging decisions. Autonomous vehicles will save lives, but their adoption could result in massive job loss. By some estimates, there are four million taxi, delivery, bus, and truck drivers in the U.S.all of whose jobs could be automated. And that doesn't even take into account the management and support staff for those jobs, nor the trickle-down impact on the gas stations, motels, retail outlets, and restaurants that rely on trucking routes. Autonomous vehicles can't vote; workers can. Policy makers will need to have serious political will to prioritize saving lives and future technology over real and perceived job loss, especially if the economic headwinds continue to blow in a populist direction. Another risk accompanies the emergence of autonomous vehicles. Remember the Wired journalist who allowed hackers to take over his carwith him in it? That hack was carried out with only a few components of the car controlled by a computer. Autonomous vehicles, by contrast, are run entirely by software. An alert driver could override the self-drive function if they realized something was going wrong, but if no one's at the wheel, there's no human safety backup. While we haven't seen any large-scale automotive hacks yet, we did see a preview of the security risks inherent to Internet-connected automobiles with last year's high-profile hack into household Internet of Things (IoT) devices. To manage the risk of a more destructive hack that could weaponize automated vehicles, significant advances in cybersecurity will be required, along with commitment from the entire connected automotive ecosystem to adopt state-of-the-art security technology. 11 The path to level 5 autonomy, or a completely autonomous vehicle that performs equally to a human, will be incremental and evolutionary, with drivers and autonomy coexisting for a long period while technology, regulations, infrastructure, and public opinions sort out the moral and ecosystem issues. Autonomous vehicles are inevitable, but fortunately for consumers, the technologies required for ubiquity will improve the safety, efficiency, and enjoyment of the ride along the waywhether there's a person, a machine or a combination of the two in the driver's seat.