Przejdź do głównej treści

Widok zawartości stron Widok zawartości stron

JU Research

Nawigacja okruszkowa Nawigacja okruszkowa

Widok zawartości stron Widok zawartości stron

Evolution or revolution? The future of AI

Evolution or revolution? The future of AI

How do we teach a machine to feel? How can we show it the difference between good and bad? What obstacles do we still need to overcome on our way to designing artificial intelligence comparable to our own? We asked these questions to Dr Błażej Sajduk from the JU Institute of Political Science and International Relations.

When we discuss the concept of artificial intelligence, we usually think about it in terms of definitions provided by mathematics or computer science. But how does it look from the perspective of a social scientist? Is it possible to adopt a definition that will satisfy researchers from different academic fields? ‘There is no one official classification that satisfies everyone. Reports on the subject sometimes contain up to nine definitions that are compatible with one another to some extent, but none of them covers every nuance of the issue. As of now, there can be no agreement, since there are several different approaches on how artificial intelligence should operate’, explained Dr Sajduk.

The birth of a revolution

Unlike in the past, when computers used simple algorithms to solve repeatable problems, in the 21st century machines became more complex due to technological advances. ‘Computers started dealing with big data. The processing power of computers started becoming more and more formidable, particularly when it came to graphics card chipsets. There also came a breakthrough in machine learning, particularly in the form of neural networks able to increase their knowledge over time. In essence, we’re transitioning from building machines dedicated to one simple task to designing complex computers capable of performing a variety of operations’, said Dr Sajduk. All of this allowed us to see artificial intelligence in a completely different light. We now try to design computers that will try to think like us.

Both lawmakers and society at large knows very well that any civilisation based on advanced technology needs engineers and programmers. But there is an ethical side to programming as well. A great example of that is the protest organised by Google employees against the Maven programme, commissioned by the US Department of Defence in order to identify potential military targets based on an analysis of human behaviour. Google staff pressured the company’s executives to drop the contract, because they didn’t want to participate in this type of project.

Designing artificial intelligence involves much more than just increasing a machine’s computing power. The most difficult problem is the one of achieving artificial sentience. As of now, we have found no way of resolving this issue. ‘A number scientists and entrepreneurs such as Stephen Hawking and Elon Musk have called for a ban or at least tight restriction on AI research, believing it to be the greatest threat to humanity. Others, like Juval Noah Harari, deem augmenting our bodies with marvels of technology to be a natural step in our evolution’, stressed Dr Sajduk.

Modern arms race

Politically, AI is also a controversial issue. ‘The Chinese “Next Generation Artificial Intelligence Development Plan” from 2017 openly talks of Chinese AI supremacy to be achieved until 2030. Russian president Vladimir Putin stated that whichever country finds a way to implement AI in its military will rule the world. Artificial intelligence can be applied in already existing weapons, such as tanks, drastically improving their effectiveness and enhancing their capabilities’, said Dr Sajduk.

The introduction of AI will also have a profound effect on the economy. Alongside 3D printing, it will fuel the coming technological revolution. Already it’s being compared to the industrial revolution that started in the late 18th century, when a great number of people started losing jobs. Should we prepare from something similar to happen to us?

‘Microsoft founding father Bill Gates proposes that we tax robots in order to make them comparably cost-intensive to humans. As for earning money, there are several ideas pertaining to universal basic income (UBI), currently tested, for instance, in the Scandinavian countries. The thinking here is that the social consequences of people not working, or working illegally, might be much more dire than those of the government paying its citizens a set amount of money to cover their costs of living. The new model of capitalism also involves moving the means of production back to developed countries. The next generation might see serious changes in the way we live, but before that, there will be a transition period’, said Dr Sajduk.

Everyday AI

We’ll also face challenges when it comes to privacy laws. Apple’s Siri and Google Assistant can be of real help in our daily lives, but first they need to collect an immense amount of data about us. This tendency toward data acquisition, resulting in ‘smart houses’ and ‘smart cities’ operated by AI. ‘Artificial intelligence will assist us in everyday activities that are currently taking a lot of time or are impossible to perform without extensive training. For instance, we already have devices that are capable of real-time translation, allowing us to communicate with people who don’t know our language’, stated Dr Sajduk.

He added: ‘Many are afraid of the possible military applications of artificial intelligence. One of the most prominent is the issue of ‘killer robots’, frequently stressed by human rights activists. Although lethal machines could be condemned or even outright banned by governments or international organisations, it’s unclear what’s going to happen in the case of robots that only stun and immobilise their targets. Things will get really complicated here’, concluded Dr Sajduk.

The challenges we will face during our technological advancement will be rife with problems with no obvious solutions, and answers might be hard to come by. After all, we don’t know how AI will behave in response to the famous trolley problem: will it simply kill one person to save five others?

Original text: www.nauka.uj.edu.pl

Polecamy również
Ambassador of Sri Lanka visits the Jagiellonian University

Ambassador of Sri Lanka visits the Jagiellonian University

Scientists continue to increase their knowledge of virus biology

Scientists continue to increase their knowledge of virus biology

Project co-authored by JU MC student awarded in the Direction: Space competition

Project co-authored by JU MC student awarded in the Direction: Space competition

Ambassador of Iceland visits the Jagiellonian University

Ambassador of Iceland visits the Jagiellonian University

Widok zawartości stron Widok zawartości stron