In the last few years, advancements in technologies have opened up promising new ways for humans to utilise robots. And although many people still fear AI and robotics by only perceiving them as what will replace humans altogether and take our jobs, a few have already realised their potential in helping the sector of personalised care.

That’s what the American computer scientist, roboticist and founder of the Interaction Lab Maja J. Matarić and her team at the University of Southern California (USC) are working towards—implementing robots in the lives of kids with autism in order to teach them social skills and help them develop.

Why children with autism? Worldwide, according to the World Health Organization (WHO), one in 160 children has an autism spectrum disorder (ASD). And while some people with ASD can live independently, others have severe disabilities and require life-long care and support.

This kind of care and support requires a lot of human-based interventions, ones that are both expensive and time-intensive. Behavioural treatment and parenting skills training programmes can reduce difficulties in communication and social behaviour, thus resulting in a positive impact on the well-being of someone with ASD. And yet, globally, access to these services and support remains inadequate. That’s where Kiwi the robot comes in.

Kiwi consists of a machine-learning model that interacts with autistic children and uses data such as dialogue and eye contact to predict whether they are engaged in a given training activity. If Kiwi understands that they’re not engaged anymore, the robot can then react and re-direct children to hold their attention on therapeutic exercises for longer. During testing, the prototype reached a 90 per cent accuracy in predicting the child’s engagement.

The study was based on the information collected after leaving Kiwi with the children for a month in their homes. The kids participating in the study regularly played space-themed math games on Kiwi’s attached tablet. The robot would then give expressive personalised feedback through a reinforcement-learning algorithm.

The game’s main purpose was to teach the kids fundamental social skills through their interactions with the robot, such as turn-taking and eye contact. With every intervention, a behavioural therapist evaluated the child’s social skills before and after in order to validate the approach for improving them.

Screen Shot spoke to Matarić about her team’s study, what assistive robots can really be useful for and how she devoted most of her career to researching their ability to help people. “We don’t yet have robots on the consumer market, at least not any robots that are truly useful, except perhaps robot vacuum cleaners like the Roomba. I hope that soon we will see actually useful robots such as robot tutors for children and robot companions for the elderly,” she explained.

In Matarić’s mind, once people actually start seeing useful robots on the consumer market, then they will be better informed about what those robots can do and will therefore know what to think of them. Until then, people will remain wary of them.

Could that be what the future holds for Kiwi? Matarić certainly hopes so. “We have developed and tested robots that helped stroke patients, healthy elderly, Alzheimer’s patients, children receiving IV injections, and of course children with autism,” she said, “so we, and others in our growing field, have demonstrated that socially assistive robots have great potential across the age and ability spectrum.”

But unlike what most people who are scared of these technologies and the endless possibilities they offer us tend to think, robots will not replace humans. Human care should not be deleted completely, but for some certain conditions such as autism, the amount of needed care greatly outstrips the time and resources many families have at their disposal. “Similarly, patients with Alzheimer’s disease often need around-the-clock care that is not affordable. These are just two examples where technology can serve to fill the gaps where human care is not available, accessible, or affordable,” explained Matarić.

It looks like we are slowly but surely accepting the presence that robots will soon hold in our lives. In Wuhan, China, a hospital ward run entirely by robots has just opened—a piece of news that we were all happy to welcome as we’re struggling to handle the impact that the coronavirus is having on so many countries. Progress in robotics is being made worldwide, each day.

Regulations and carefulness will be essential for everything to go smoothly, but the hope that such robots will become an affordable companion for children with autism is enough to get me excited about the near future.

The post Robots aren’t taking over the world, they’re teaching kids with autism social skills appeared first on Artificial Intelligence.

Machine learning is becoming more prevalent in the modern classroom as teachers make the shift to blended learning. Students often enjoy the more modernized version of lessons that allow them to learn at their own pace. Even teachers can enjoy the more personalized aspect of machine learning. However, many are beginning to wonder if everything about these machines is truly beneficial for our students. It seems that there are just as many limitations to learning this way as there are benefits.

It’s time to take an objective look at the real advantages and disadvantages of machine learning. When we can face the truth about what this holds for our children, we may be able to better balance our expectations.

Benefits of Machine Learning

It’s hard to overlook personalized education as the most prominent advantage of machine learning. Students can work on material that suits their unique capabilities and advance to more difficult content when they are ready. Not only do students benefit from this unique approach to learning, but teachers can save a great deal of time. They no longer have to create lesson plans that cater to students of all abilities and grade levels.

The advantage of machine learning also shows up in the automatic grading system. This presents the opportunity for completely unbiased grading that can’t be influenced by the teacher’s relationship with any student. It saves teachers time and gives a more realistic overview of a child’s achievements in school.

Machine learning can also help educators to look toward the future. By analyzing their data in the system, patterns will quickly emerge that show where the student’s primary weakness is and if they will need more help. If the issue is academic, the computers can provide a form of virtual tutoring. However, it can also indicate if a student is at risk of dropping out or receiving greater disciplinary action.

Limitations of Machine Learning

The major downside to machine learning is that we are taking personal interaction away from the students. This can dramatically impact their ability to make friends and present themselves well in the workplace over the years ahead. Social skills still need to be emphasized even while using machine learning.

As helpful as it can be to allow the computer to grade student papers, it isn’t always going to be effective. Educators will still need to plan to grade essays and other items the old-fashioned way. Computers lack the ability to assess items that don’t have a specific technical requirement associated with them. They can easily grade multiple choice but struggle with the more time-consuming essay assignments.

Of course, many schools are evaluating the advantages of machine learning due to the inherent cost. It can be quite expensive to purchase all new equipment and programs to make personalized learning an effective solution for educators and students. This can be one of the biggest limitations to machine learning.

The Final Answer

Machine learning definitely presents a number of advantageous features for students and their teachers. The possibility of a personalized education that can save time for teachers and help to predict a student’s future success is invaluable. However, we still have a long way to go. Educators won’t want to jeopardize the social skills of their students, and they must consider if their school district can afford the equipment. Schools will have to decide how to balance out the features that mean the most to them and how they can afford to implement them.

The post THE BENEFITS AND LIMITATIONS OF MACHINE LEARNING IN EDUCATION appeared first on Artificial Intelligence.

Self-learning machines and smart robots are the basis for future technologies in many industries. AI gives computers humanoid and cognitive abilities that can lead to an increase in productivity. In addition, machine learning (ML) systems are employed to assess problems and needs and resolve them accordingly.

In recent years, extensive progress has been made with artificial intelligence: Watson, Siri, and deep learning are just a few examples which show how AI systems can now perform services which can be classified as intelligent and creative. Companies that want to optimise their business or save costs will hardly be able to dispense with artificial intelligence. A computer, much like a human, saves knowledge gleaned from experiences and can then independently derive solutions – based on algorithms – to new and unknown problems. To do this, a computer program analyses samples and, using self-learning algorithms, tries to recognise certain patterns and regularities in the data.

AI systems are, without a doubt, extremely useful
The world is becoming ever more complex, meaning that a well thought-out and wise use of our human resources is required. This is where high-quality computer systems can help. It also applies to applications that require intelligence. The downside of AI is that the possibility of intelligent machines is something that scares many people who argue that intelligence is something unique, something that sets Homo sapiens apart from mere machines.

The quest for an artificial copy of man and the issues associated with this are not entirely new. Our ancestors already invested much thought in the reproduction and imitation of thinking.

Economists expect that this technology will greatly increase prosperity. Previous studies have been restricted to investigating the impact of intelligent automation, for example, when machines in networked factories communicate with each other or when autonomous vehicles, transport goods or people without drivers. This view, however, neglects other growth-stimulating effects of intelligent systems: artificial intelligence will support people in their work and make them more productive.

Economic growth through AI
According to a recent study by the consulting firm Accenture, artificial intelligence will give the German economy a huge growth spurt. Many work processes will become more efficient and staff will be able to focus more on tasks with a high value added, as machines and programs take care of many routine activities formerly dealt with by not only assembly line workers, but also knowledge workers. According to a different study by McKinsey, the global market for AI services has an annual growth rate of up to 25% and is projected to be worth around 125 billion euros by the year 2025.

Machine learning is becoming more and more mainstream. While self-learning programs were only to be found in universities, research institutions, and some technology companies up until a few years ago, nowadays they are finding their way into ordinary products and solutions. Our everyday life and our business lives are increasingly being determined by intelligent programs that learn from data and extrapolate from what they have learnt.
Speech recognition on mobile phones, for example, is essentially controlled by machine-learning algorithms – much like the spam filters in PCs and notebooks or face detection in the management of photos. We are often in contact with learning systems without being aware of it – such as personalised online advertising. More and more companies are recognising the value of machine learning when it comes to optimising their business and saving costs.

What work is left for the people?
Although robots can take over communications, computing, and thinking, there will be limits even here. There will be still be the need for highly qualified professionals, such as engineers, architects, or judges. Also tasks, whose dexterity remains beyond that of robot fingers will remain for the foreseeable future.
Machines cannot replace the emotional intelligence of a person. Occupations that require social skills or creativity or represent a high-quality personal service cannot be substituted by AI.
Certainly, some professions will disappear completely and, in many areas, the requirements will change greatly. However, there is no need to panic just yet.

Flexibility is needed
Tech companies such as Apple and Microsoft want to automate as many working processes as possible with learning machines.
The employees of tomorrow must be more flexible, because in the future man will have to adapt to the computer and not vice versa.

Employees will have to adjust to more flexible working practices: a fixation on permanent locations and times is often no longer required. This implies more flexibility and freedom on the one hand, but on the other hand, that work and personal time will intermingle.
In addition to the technical expertise, specialist and managerial staff must bring a more in-depth process knowledge and have a higher willingness to undergo independent and ongoing training in the appropriate technologies. What is more, a good understanding of all security-relevant questions relating to IT technology and legal security will be a basic requirement.

New careers in IT
According to Bernhard Rohleder, Managing Director of BITKOM, new, exciting and challenging jobs will be created. However, they require a good education and offer much creative freedom and responsibility.

The number of workers with low-grade qualifications will fall, but staff with correspondingly high, mainly digital skills are now already increasingly in demand. The number of employees in the IT industry has been increasing continuously for years, as do the number of vacancies in this sector that are difficult to fill.
And there’s more: even if artificial intelligence means that we are talking about machine learning, a harmonious combination of machine and human intelligence working as a single unit is required to develop theories and to confront new challenges. Working on the basis of the existing data, it is possible to develop company-specific business ideas, whether in the form of products, processes, or new data-based business models.

The post Artificial intelligence and the future of human labour appeared first on Artificial Intelligence.