Preparing a case is a complex task for lawyers. They have to examine laws and judgments, collect evidence and develop strategies that are tailored to the presiding judge. “Everlaw” is a software that helps American lawyers with these tasks. It is an artificial intelligence (AI) system that can read documents – and locate those that are relevant to the case. “Lex Machina”, another AI software, also provides profiles of judges. This AI knows what decisions the judges tend to make and predicts how long a case will take to resolve and what the outcome could be. It also knows the weaknesses of the opposing lawyers.
Such technologies, which reduce the amount of research work, pose a big problem for legal assistants: they are becoming increasingly superfluous. In the USA it has become difficult even for law graduates to find entry jobs in law firms, because typical beginner jobs are done more efficiently by computers.
Decades ago, automation drastically changed production and agriculture – now it is reaching the service sector. Machines are able to relieve us mentally. This will affect all areas of life, and it is not foreseeable what we will face in the coming years.
The AI initially met with the most approval in those sectors in which statistics and figures are juggled, especially in banks and insurance companies. This happened almost unnoticed – despite the fact that automation brought enormous organisational changes and cost many jobs in customer service. Gradually, automation – also without much fuss – moved into the private lives of many people. Artificial intelligence helps daily in the internet with search inquiries, filters messages for us and recommends products in online shops, which we can buy.
Many people use digital assistants such as Siri and Google Assistant to manage their appointments, find restaurants or plan trips. With Alexa from Amazon and the Google speaker, the assistants are also physically present in the home – answering questions, entertaining children, playing music or movies on demand, and they control the intelligent house by regulating the heating, operating washing machines, or monitoring the entire house when you’re away. Robots like “Jibo” by social robotics expert Cynthia Breazeal have been specially developed to master social conversation and, using face recognition, to recognize the people they are assisting and turn around to when they are called.
When we think of robots today, we often still have the human-like robots from science fiction films in our minds and overlook the fact that we have long since been surrounded by real, less spectacular-looking robots and artificial intelligence without consciously perceiving it. The increasing ability of Artificial Intelligence to cope with everyday human life and the tacit acceptance of such systems make it attractive for many service professions – especially those in which there are repetitive activities where humans are prone to error.
This is most evident in road traffic. Not only Google is testing – now under the name Waymo – a self-propelled car, but almost all other car manufacturers have also jumped on the bandwagon. In Germany, several test tracks are being prepared in various federal states. One of the better known is the A9 between Munich and Nuremberg, where Audi has tested vehicles from the Piloted Driving project in recent years. Daimler and BMW are also investing in such autonomous technology – they intend to bring the first self-propelled models onto the market in the next five years. Previously, it had been assumed that this would take about 15 to 20 years, but the necessary sensor technology has now made considerable progress.
In Bad Birnbach, Bavaria, Deutsche Bahn tests a self-propelled bus in local traffic. Deutsche Bahn’s vision is that in future customers will order a small bus to their front door with their smartphone, which will then take them to the destination of their choice. Lines and stops would thus be superfluous. Autonomous shuttles will also be tested at Frankfurt Airport and soon at the Charité Clinic in Berlin. The railway is also preparing the first autonomous trains – initially freight trains.
The automation of road traffic makes perfect sense. The robots should ensure environmentally friendly, fast and safe traffic. According to a study by the Lawrence Berkeley National Laboratory, electrically powered autonomous vehicles could emit about 90% less greenhouse gases at optimum capacity. Researchers also assume that the number of accidents will decrease considerably – robotic vehicles react faster, are never tired or drunk. But of course automation has consequences: The AI expert Wolfgang Ertel writes in the book “Grundkurs Künstliche Intelligenz” (Basic Artificial Intelligence Course) that it is almost certain that there will be no more taxi drivers in developed countries from around 2030. The same is likely to apply to drivers of rail vehicles and not least to pilots – automation is also underway in the airspace.
The situation is no different in non-personal logistics. DHL is testing robots that automatically accompany postmen and help them to deliver their mail. In the process, they learn how to cope on the street on their own. DHL and Amazon are also testing delivery drones, which are initially to transport urgent deliveries – especially blood and medicines – to hospitals. There is no question that in future parcels will also be delivered by robots and drones.
Ocado, a British online supermarket, has opened a department store in Hampshire that will be operated entirely by robots. Around 1,000 of the smart machines swarm back and forth on an area the size of several football pitches. They manage to process an order in five minutes, which takes two hours for human staff in the company’s other warehouses. What the machines have not been able to do so far is package sensitive vegetables without damaging them.
The Henna Na Hotel in Japan will also operate automatically. The hotel staff consists of 140 robots who take care of guests in a total of 100 rooms without having to go to the toilet or have lunch. At the reception, for example, a robot looks like a prehistoric Raptor, but speaks Japanese, Korean and English. A self-propelled transport robot takes the guests’ suitcases to a robot arm that lifts them into compartments to stow them safely.
A cute robot doll on the bedside table in the hotel rooms verbally accepts the wishes of the guests. A room service drone is currently being tested. Only the beds have to be made by human employees. However, this is not likely to be the case for much longer. The Belgian supermarket chain Ahold Delhaize wants to robotise all cleaning and cleaning work in five years. Like a science fiction film, the doors close at the close of business and a robotic army comes to life.
2012 the American company, Momentum Machines, introduced a robot that can prepare 400 hamburgers in one hour – from cutting tomatoes to roasting meat to folding and packing. The technology is still under development, but the American fast food company Caliburger is already using the simpler Flippy robot, which roasts and turns meat burgers. The British company Moley recently introduced a completely automated kitchen in which two robot arms handle from above over the stove and prepare entire menus – not faster than humans, but more convenient.
Of course, artificial intelligence does not need hands to replace people. In the service sector, administrative activities can also be automated – and thus the administration and the secretariat. “Voicera”, for example, is a startup that developed the AI assistant Eva. Eva takes over the task of taking notes in office meetings. She creates a summary from the texts after the meeting and sends it to the participants while they are still packing up their documents.
The assistants “Amy” of the company x.ai takes care of appointments. All the user has to do is to put Amy in “CC:” in his e-mail and to point out that she will get in touch with the addressee, like this: “Hello, would you like to have coffee tomorrow? If so, Amy will contact you about the time.” Amy understands natural language. She reads the mail and contacts the conversation partner to negotiate the appointment. She has access to the user’s calendar. If the person writing replies, “I don’t have time tomorrow, better would be the day after tomorrow at half past three”, Amy can also understand this and adjust the appointment.
The “Mya” software is even more sophisticated. This chatbot takes over the organization of job interviews. As if one would chat with a friend, the Chatbot replies to for applicants. It finds out by clever questions whether the applicants fulfill the minimum qualification for the advertised place and supports them in sending in their complete documents. It also answers questions about payment and organizes the interview.
And so it goes with the automation further and further. Even creative activities are less and less safe from machines. There is software that can compose music, and robots that can apply the style of great artists to painting. Literary scholars use AI systems to analyze literary texts. The advantage: In contrast to a professor, the software can read all relevant texts, keep them in memory and identify content-related or linguistic influences that people overlook with their limited reading time. It is estimated that Artificial Intelligence will also outperform humans in tasks such as translating texts, writing simple essays and journalistic articles over the next ten years.
However, entrepreneurs do not like to talk about automation. And if they do, their standard answer is that automation should support people, not replace them. It is intended to make people’s work easier so that they can concentrate on more complex, fulfilling tasks. This is partly true, but there is no question that entrepreneurs are also looking to cut staff costs. A robot costs several cents per hour over the years. From a purely financial point of view, you can replace a single person with a double-digit number of robots or AI software, depending on the task at hand. The McKinsey Global Institute has published a study that estimates that 800 million people worldwide will lose their jobs in machines by 2030.
That has consequences. When Danish hospitals recently introduced robots to provide doctors with material, employees started sabotaging the robots by leaving faeces to clean up, the New Yorker reports. On his cover, the magazine showed several robots strolling through a street as passers-by, inflicting a few cents on a homeless person.
Most people, provided they are not affected by the loss of their job, otherwise accept automation because they benefit from a better service – even if they have to provide insight into their data, for example. In China, if people want to buy a car, they can use their smartphone to send a short message to their bank’s chatbot in order to take out a loan. The software immediately checks whether it can be approved or not – and sends the approval or rejection within seconds. This is possible because companies in China are allowed to collect and evaluate almost all of their customers’ data – they have a good picture of their creditworthiness.
China may seem extreme, but the West also gives in on data protection. IBM’s Watson AI system is also in use in banks and helps to check whether customers’ insurance policies cover everything they want or have gaps. The same system is used in medicine in some countries: It scans immense amounts of data, including patient records and research results. It records the symptoms of current patients and suggests diagnoses to the doctor. In fact, Watson has already been able to identify rare diseases that have puzzled doctors. What can make data protectors uncomfortable is sometimes a blessing for patients. Such contradictions often arise during automation.
Nowhere is automation as controversial as in the areas of care and education. Robots are already active in retirement homes, such as “Pepper” from the French company Aldebaran Robotics. It has been developed to respond to different voices and facial expressions of seniors. This allows him to talk to people, play with them and serve them drinks or snacks. The robot “Paro” is even more widespread: It is a kind of pet replacement, reacts to touch, blinks its eyes and hums contentedly. In this way it demonstrably demands the well-being and thus the ability to concentrate of people suffering from dementia. The Milo robot, which was developed to help autistic children recognize emotions, seems similarly effective. The manufacturer advertises that Milo is always in a good mood, never frustrated or tired. The credo is that machines are the more patient carers and educators.
The question we must ask ourselves is not what machines can do, but what we want to leave to them.
This article was published in a modified form in the journal “BuB – Forum Bibliothek und Information”.