STAznanost

No need to fear robots or artificial intelligence

Ljubljana, 24 December - There is no need to fear robots or artificial intelligence (AI): the former are at least for now not capable enough to be dangerous to people, and the latter does not even exist as such. AI is a series of steps to tackle complex problems, steps simple enough that someone with a high school knowledge of maths can master. These were just some of the points raised at an online debate hosted by the web portal STAznanost, where the participants noted that development has recently been rapid in both fields, yet not as rapid as we would have wanted or expected.

According to Marko Grobelnik, who co-leads the Department for Artificial Intelligence at the Jožef Stefan Institute, AI first needs to be demystified and included in curricula at all levels of education. "The technologies are indeed not very complex. These are mainly a series of procedures that actually someone with a high school level knowledge of maths can master."

Marko Munih, the head of the Laboratory of Robotics at the Ljubljana Faculty of Electrical Engineering, and Aleš Ude from the Department of Automatics, Biocybernetics, and Robotics at the Jožef Stefan Institute, believe these topics should absolutely be introduced into every "pore of education" to find the talents of the future who would continue their career in research or at companies.

A lot has been done in education in robotics in the past, as the first textbooks in Slovenian as well as in English were published in Slovenia in the 1980s. "We have competent people in this field. If this hadn't been so, we probably would not be talking about this today," said Orest Jarh from the Technical Museum of Slovenia, which currently hosts the exhibition Robot.si, which traces the development of Slovenian robotics, especially in industry.

How did we get to the point where we are today?

The Technical Museum put on the exhibition to mark 100 years since the first mention of the word robot in the 1920 play R.U.R. by Czech writer Karel Čapek. Back then, robots as we know them today did not even exist. Automated machines existed of course, but under today's definition a robot is capable of conducting tasks or a series of tasks that are controlled by a computer and must also be capable of conducting at least three different types of movement.

The first real robots were Elmer and Elsie, made by William Grey Walter the late 1940s. They did not know much, but they were able to "return to their house to recharge," said Jarh. Elmer and Elsie were not digital robots yet. They were analogue.

The real breakthrough came in the mid-1950s, when George Devol made the Unimate robot in the US. The robot was first launched at the factory of the car giant General Motors in 1961.

This is the period when AI was conceived - at least the algorithms that are still used today, for example neural networks. However, a key component was still missing - computer power that would allow for such demanding operations, said Grobelnik.

Since robots did not need such computer power, the development of this area continued - soon Japan was in the lead globally. Slovenia followed quickly, with the Jožef Stefan Institute and the Ljubljana Faculty of Electrical Engineering starting to deal with robotics in the 1970s; in the 1980s the very first Slovenian-made robot, Goro 1, came to life, developed and manufactured by the Jožef Stefan Institute and the home appliance maker Gorenje.

This cooperation continued until the 1990s, when robots became easier to import, and experts and researchers started focussing on the computer part of robotics and its application.

As Jarh put it, the development of robotics in the last 100 years was very interesting. "We had writers and artists who started thinking about robots before we actually had robots. They were humanoids but these were no longer in the picture when engineers took over. They used or preserved only what was needed. A factory robot does not need two legs."

In general, robotics development followed the development of other areas. According to Munih, at first milestones in robotics followed breakthroughs in electrotechnics, then came the development of microprocessors etc.

Progress in robotic driven by industry, but development moving towards man-robot partnership

Robotics has been driven mainly by industry throughout this time, Munih said, noting that the return on investment was fastest in industry and created profit. But today, researchers are increasingly focussing on other areas of robotics, said Ude.

"Of course, industrial robotics is still important but it is not the only topic. Currently, the most pressing issue is how to move robots from factories into a natural environment, where they will work together with people. (...) In the distant future, when our know-how is better, perhaps robots could become some kind of natural partners to humans."

At this point, autonomous vehicles probably receive the most attention, as this field is definitely in for a major breakthrough in the near future. Another interesting area is medicine and medical care.

"We're not as far as with autonomous vehicles there. But in the long-term more robotics can be expected in medicine as well," said Ude, stressing that at first these would not be humanoid system but simple robots, like for example automatic robot vacuum cleaners of the kind already used in homes.

In the long-term, Ude believes development will focus on hardware, materials that robots are made of. "Present-day robots are still very metallic, solid. Soft robotics - manufacturing of robots that are more pleasant to work with, to the touch, and have algorithms to work better with humans."

The more complex the systems, the more complex the ethical issues

Ude pointed out that robotic systems which are more complex also raise more complex ethical issues - for example the well-known dilemma of whom an autonomous vehicle would protect as a priority if a collision cannot be avoided - the driver or the pedestrian.

Many stakeholders emerged in the AI arena at the international level in the last couple of years, including UNESCO, the UN, the OECD, Council of Europe, the European Commission, and as part of these efforts those also the Jožef Stefan Institute and its IRCAI - International Research Centre On Artificial Intelligence, which operates under UNESCO's aegis in Ljubljana.

According to Grobelnik, the main goal of their efforts at the moment is to regulate AI technology so that it would not cause harm on the micro level or on the global scale.

But as Ude said, "people are often afraid of robots. However, I think this fear is completely unjustified - not least because present-day robots are not capable enough yet to be a peril for humans".

After all, the algorithms and AI that are used by robots are developed by people.