Digital Humanities is an emerging research area. Wikipedia says that “Digital Humanities (DH) is an area of scholarly activity at the intersection of computing or digital technologies and the disciplines of the humanities.” This statement doesn’t fully reveal anything clear or concrete. Whatever the definition, it is important to develop more services where humans and machines work together better. In most cases, AI is not an independent machine that handles all tasks but a tool to help people. That’s why we need research and development to get this interaction working better.
Applying AI in the realm of education
I have followed a university research project where software and AI can better support teachers and students in language studies. In this project, artificial intelligence, language technology, applied linguistics and educational data science are used together to better tailor the learning experience for each student. A student can, for example, take articles from their interest area, and software prepares language exercises for them based on the articles themselves or from their skill level and earlier exercise results.
In this way, the students are more motivated to work with topics and content they are interested in or need to practice. At the same time, when the system knows a student’s own and other similar students’ skill level, strengths and weaknesses, it can support the student like a personal teacher. The objective is not to replace language teachers but to support teaching and the student’s self-studies. This solution is based on top-level science, where supercomputers have been used to analyze masses of data to develop models.
There has been software to study new languages around for years, but most are geared for beginners and tourists who want to order a beer and a steak. Some of these services also promise that you will speak a new language fluently in six weeks. We all know that it will not happen, but sometimes we hope it could happen and still pay for those services.
The challenge is to have solutions and software that can help people when they want to learn a language seriously, both spoken and written. This is an excellent example where AI can work with humans (i.e., teachers and students) to support them to achieve better results. In practice, it means a combination of computer, linguistic and education sciences.
Making human-machine interface meaningful
There are many other areas where computer science, AI, and social sciences need more interaction. AI ethics is another hot research area. As I wrote earlier, it is a complex area and still has more opinions than systematic research and development. It is, however, an excellent example of how important it is to focus on the interaction between machines and humans. As in the case of learning a new language, it is about getting machines and humans to work together. Furthermore, it is not only about machines and their functions, but how machines impact people and how people learn to use machines and change their behavior when working with machines.
From science fiction and some business plans, we hear about super-smart, independent robots. We read daily about new AI solutions and how they replace human work. At the same time, many skeptics are claiming that robots are not even close to mimicking human beings.
Guilty Robots, Happy Dogs is a philosophical and psychological book about robots, pets, and human beings. People often want to humanize other entities and think they behave and think like human beings. When your dog does something, it’s easy to explain its behavior as person-like. But there is no evidence that animals think or behave based on the same model, consciousness, and feelings as humans.
We quickly make the same mistake with AI and machines. It is easy to think AI is still the same as those movie robots that walk, feel and behave like people. Then it’s easy to conclude that they are still far behind human beings. But that perception is quickly misleading. Machine learning and AI emerge from data analytics. Those machines can collect a lot of data from many sources, learning from that data and then make decisions and take actions based on the data. These machines are then integrated into other systems and devices to guide and control them.
The machines help people in operations that they can handle much better than people, such as analyzing millions of data points each second and making conclusions based on them. If we compare AI and humans, it may seem that AI is born with a purpose, and humans spend most of their lives trying to find one.
Building better human-machine frameworks: paths forward
We can no longer ignore the role of machines and AI in business, from manufacturing and car driving to education, health care, finance, legal work, or decision-making support. It also means people must constantly learn to utilize machines better and interact with them. But the systems must also be developed to better ‘understand’ and support humans, an area we definitely need to research further.
When we think about solutions to help people learn languages, avoid ethical problems with AI decisions or get machines to help people with important decisions in life and business. AI development is not about developing independent super-robots but software, solutions and machines that better work with people. It is about getting more data, optimizing algorithms, and understanding how people and machines can interact and support each other. Computer science skills are not enough; we also require humanities and social science knowledge.