AI is the acronym of the moment. And it seems that in 2020 it will advance in fields as important as education, logistics, healthcare, transport and even the military. And yes, our privacy will be affected.
In 1997, IBM’s supercomputer called Deep Blue was the first to defeat world champion Garry Kasparov at chess. In little more than 20 years, artificial intelligence has developed in a multitude of fields and has revolutionized citizens’ lives. In 2011, virtual assistant apps such as Siri, Now and Cortana transformed the technology market; in 2013, Google created an artificial brain; 2013 brought the machine learning boom; in 2015, people began to talk about autonomous cars and Alphago became the first machine to beat a professional player of the game Go. In 2017, image-processing algorithms that improve the tracking of criminals through facial composites came to the fore, and this year AI has consolidated its role in recruitment. But what awaits us in 2020?
- Medicine: automatic diagnoses
AI will help healthcare professionals diagnose diseases automatically. A machine can give them clues so that they do not waste time, and reduce doctors’ workload. The detection of tumors in MRI and CT scan images is already being carried out in some hospitals in the US. And they outperform 3D images, which present many slices, as the machine is able to rule out all those where there is no tumor. “The reports from IBM (the creators of this system) claim that the algorithm exceeds the accuracy of a human. However, the system was brought to Europe to diagnose other types of diseases and did not achieve optimal results, so improvements will have to be made in 2020,” says Jordi Casas, director of the master’s degree in multimedia applications at the Universitat Oberta de Catalunya (UOC) and an AI expert.
Innovations in medicine are, without a doubt, the greatest that artificial intelligence has brought, according to researcher Humberto Bustince. We are able to classify, with all the data, all the patients who arrive at the emergency room,” he explains, “and variables can be handled that used to escape doctors because of the sheer volume of data involved.” It is also capable of allowing the machine to read the mind “so that they can tell you what they would prefer to eat without needing to move or open their mouth.”
- Logistics: Your supermarket knows what you are going to want
Supermarkets, large retail outlets and small shops are preparing for a new way of consuming. These types of establishments are beginning to implement demand forecasting systems. “They are starting to know which products a user buys in each store with machine learning, and they prevent delivery trucks from traveling in vain while never running out of stock,” Casas explains. To find out which products they need, they start from previous data, from what they have sold and on which specific dates, whether it was Christmas or Black Friday. “AI systems accumulate data and then the algorithms learn from the previous data, renewing their effectiveness year after year,” Casas adds.
Evidently, leadership in this regard comes from China: there, cashiers are disappearing in favor of automatic payments via mobile phone. Carts that follow the customer or conveyor belts with items are already part of the landscape.
- And they also know how much energy you are going to consume
AI will help the most vulnerable families. There is a pilot program for detecting households that have greater difficulties coping with water or gas bills. “A pattern is established with the homes that in one month or several months have had difficulties paying due to delays or outright non-payment; buildings with low energy ratings are examined; the cases of single-parent families who may have lower incomes are analyzed; and in this way, social services can focus on certain areas of specific neighborhoods,” says Casas. “In addition, electricity companies are working to generate enough energy to meet demand without producing too much, so as not to waste it,” he adds.
- Paid audiovisual platforms: the science of influence
In 2011, Netflix had 21.5 million subscribed users. Seven years later, this figure has multiplied by seven: Netflix has 151.6 million and HBO totals 142. The explosion of viewers on video-on-demand platforms (SVOD, for subscription video-on-demand) has not come alone; the awards have arrived at the same time: in 2019, HBO took home 34 Emmy awards and Netflix, 27.
For years now, when a user signs up for Netflix, HBO or Spotify and begins selecting films, series or songs, the algorithm stores the user’s preferences and recommends new content based on those preferences. What we do not know is what lies behind the science of influence. “Many of these platforms design different posters for their own series and, depending on whether the user’s preferences are dramas, comedies or horror series, they show one poster or another so that the user feels drawn to play it, whether it is to their taste or not,” said Keith Dear, a member of the United Kingdom air force’s innovation department, during a talk at the AI and Big Data congress Big Things. This has given rise to scandals. Such as when it was detected that Netflix’s black customers received posters featuring the black actors who appeared in the films, even though their roles were marginal.
- Military practice: artificial generals.
It sounds like science fiction, but next year it will be a reality. “AI is going to alter combat decision-making in the army,” Dear asserts. The precision of satellites and the improvement of algorithms will mean that sending soldiers to a given place, with specific coordinates and taking one specific route, means a higher probability of survival than taking another path. “Simply by analyzing other battles and the ways other soldiers have acted, an algorithm will be able to design better strategies than humans themselves,” Dear notes. “However, we are concerned because there is a major ethical implication: it is possible for a machine to be better at recognizing the enemy than a human in the same position,” he adds.
- Education: a robot teacher
The current computer systems within most universities store students’ academic data, such as their grades or previous studies, but they do not store other data such as the time they devote to studying, the academic resources they consume or simply which competencies they have acquired in previous courses.
However, some faculties are beginning to store less academic data on students with AI systems. “In some institutions, coaching systems are being implemented to recommend to students which studies they should pursue or which subjects they should enroll in based on their preferences and, in addition, to solve administrative problems they may have,” says David Bañeres, a researcher at the eLearn Center of the Universitat Oberta de Catalunya (UOC). Systems for detecting at-risk students are also being introduced. “With the grades and the actions students take in their courses, it is possible to predict whether the student is going to pass a subject or not,” Bañeres states.
Adaptive learning systems are the most complex. The goal is to emulate the figure of a teacher in every sense. That is, answering questions, recommending learning resources or activities to practice a specific topic, and automatically assessing the activities completed by students. “Until now, universities have been developing this system in separate parts. For example, some use a chatbot in the classroom to answer the repetitive questions students ask every semester, or self-assessment systems depending on the subject, but the goal is to implement a complete program,” he adds.
Progress is also being made in AI-based teaching for people with learning difficulties caused by conditions such as Down syndrome, autism or Asperger’s
- Air transport: agility versus privacy
The second half of the 20th century meant the democratization of tourism in the Western world. In the 21st, new challenges arise: making it reach everyone in a sustainable way — that is, to scale, in startup jargon — and to break with what came before, the well-worn disruption. While the sector’s classic players join the wave of transformation by trusting in alliances with startups, the latter propose more striking and experimental models.
Iberia has implemented a facial recognition pilot project in Terminal 4 of the Adolfo Suárez Madrid-Barajas airport to speed up boarding. This proposal will allow users to identify themselves at the general security checkpoint, in the fast track and at the boarding gate with their biometric profile, without needing to show their ID card or passport. These projects speed up travelers’ procedures in exchange for customers’ privacy, since it is essential for the company to store the user’s face in a database.
originally published By
