Digital First. The Threats and Hopes Pinned on Digitisation

The term “artificial intelligence” was coined by American computer scientist John McCarthy in 1956 at a scientific conference at Dartmouth. He defined it as “the science and engineering of making intelligent machines.” Today’s AI research draws on the theses and accomplishments of computer science, psychology, philosophy, cognitive science, linguistics, economics, probability, and logic. As a result, AI is now defined in two distinct ways. Ontic – as a distinct entity endowed with self-awareness and self-learning capabilities, and praxeological – as a technology that supports humans in areas requiring repetition and precision.

A mere quarter of a century after McCarthy, in 1981, Stanislaw Lem published a book entitled “Golem XIV”, a record of lectures by a fictional advanced digital machine created in the 21st century that developed artificial intelligence. “It is not true that we created the Golem. It was created by the laws inherent in the material world, and our role was limited to the fact that we were able to observe them”, said the writer. The protagonist of his book, a supercomputer endowed with consciousness, delivered a scathing critique of our culture, much to the surprise of its creators. Its statements are a dialogue between man and Reason, which is already beyond our powers of cognition.

“The Golem has no personality or character. In fact, it can concoct any personality it wishes for itself. The two sentences above are not mutually exclusive, but form a vicious circle: for we cannot resolve the dilemma of whether that which creates different personalities is itself a personality. (...) Deprived of the affective centres typical for human beings, the Golem is not capable of displaying feelings. It can only imitate emotional states, because that makes it easier to formulate an expression that will reach its addressees as precisely as possible”, Lem wrote.

And further: “We have only one quality in common, and that is curiosity, which nothing can suppress, much less destroy. It is the only point where we meet”.

Should we not take this visionary writer’s message as a warning?

Necessity and doubt

It would seem that the general opinion about the development of AI work is predominantly enthusiastic. In 2019, the Research Institute of the Scientific and Academic Computer Network (NASK) published the results of a survey conducted among Polish internet users. When asked about the benefits of widespread deployment of AI, they stated that it would improve comfort of life (34%). Almost every third respondent thought that it would contribute to reducing the number of accidents at work (32.7%), as well as generate products and services that are more suited to customer needs (30.3%). The smallest benefits were expected to materialise in the socio-economic field. Only 6% believed that unemployment levels would fall.

Digitisation of many areas of our lives is seen as a necessity, although, in addition to hopes of solving the problems that plague the modern world, it also raises doubts about the direction of its development. International cooperation is essential, hence periodic meetings of experts to agree on common standards.

Two years ago, the European Commission published the “White Paper on Artificial Intelligence – A European approach to excellence and trust”. The document is part of the European Digital Strategy, though it is not a law, but rather a set of proposals that outline the future direction of EU regulations and initiatives in the area of artificial intelligence.

Regulating the undefined

The EC assumes that regulations should take into account the degree of risk associated with AI-based solutions. Therefore, the “white paper” states that technology based on artificial intelligence should be strictly and transparently regulated by humans, especially in high-risk sectors, which include healthcare, transportation, law enforcement and the legal system. The white paper also stipulates that systems using artificial intelligence should be subject to state oversight and control.

On 22 November 2019, a report by the Expert Group on Liability and New Technologies (“NTF”) on legal liability for damages arising from artificial intelligence and other emerging digital technologies was published. The authors of the “Liability for Artificial Intelligence and other emerging digital technologies” expert report emphasise that current legislation in the Member States and throughout the European Union inadequately and inefficiently regulates the issue of legal liability for artificial intelligence and new digital technologies.

According to the authors of the report, the unique characteristics of AI and new technologies – their complexity, the possibility of changing the essence of operation after modification, limited predictability and vulnerability to security threats – require significant changes in the law. Particularly noteworthy is the expert conclusion on giving legal personality to artificial intelligence. According to the authors of the report, this is not desirable, because the damage caused by new technologies should be attributable to the responsible persons or institutions.

Ants and human values

However, this is only one side of the coin. According to the aforementioned NASK Research Institute report, almost half of the respondents (44.8%) stated that technology based on artificial intelligence will become independent from human control, and every third respondent believed that the consequences of this will be unfavourable for humans. This concern was shared by late Professor Stephen Hawking, a British theoretical physicist specialising in astrophysics, who passed away three years ago. He posited that if we succeed in creating a true artificial intelligence, it will either be the best or the worst thing in human history.

Hawking emphasised that the real risk of AI is not that it will be evil or malicious, but will stem from the competencies with which it would be equipped. “It’s going to be great at achieving goals, but if those goals aren’t consistent with our own, then we’re going to have a problem,” he said at a conference in June 2016. “I’m guessing you’re not someone who hates ants and tramples them in a fit of anger whenever you have the opportunity to do so”, Hawking noted, responding to a question from a lecturer. “But let’s imagine you’re in charge of building a hydroelectric plant and there’s an ant colony in a place that’s going to be flooded – well, tough luck for the ants. We should not let humans become such ants”, said the acclaimed scientist.

That’s why, together with innovators such as Elon Musk, Professor Hawking helped develop 23 principles to protect us from the evils of artificial intelligence. One of those principles states that highly autonomous AI systems should be designed in such a way that their goals and the way they operate are consistent with human values. These have been dubbed the “Asilomar Principles” after the name of a conference held in 2017 in California.

If we are careful to follow them, we should avoid disaster and more . Perhaps AI can help us develop a different economic model than the current one based on continuous growth, which is sometimes called the cancer logic. It is also possible that it might finally solve the mathematical problem of Bernhard Riemann’s hypothesis, which has been waiting to be resolved since the mid-19th century. Maybe it will fulfil Roger Penrose’s dream of developing a theory that would reconcile the contradictions of quantum mechanics and the theory of relativity. The challenges for artificial intelligence are many, and we can only hope that in a world ruled by algorithms, the vision found in “Golem XIV” does not turn out to be prophetic. “Dr Popp believes that the Golem does not deal with humans at all – because it knows that it cannot learn anything relevant from them”, Stanislaw Lem wrote. So if AI only learns from its own mistakes, the cost that we will have to pay will be the crucial issue. And it is not just a matter of money. ©℗

UN Digital Summit – Internet Governance Forum

“This is already the 16th edition and we have been trying to host this event for many years by organising local chapters. The theme of the summit is ‘Internet United’, in reference to an internet which is united, accessible to all and provides more opportunities. The last year showed us how important the internet is; without it, we would not have had access to many services or have been able to handle many administrative matters for work or business”, says Krzysztof Szubert, Republic of Poland’s Plenipotentiary for the UN Digital Summit – IGF 2021.

The summit will last one week. Over 300 debates and panels of various kinds are scheduled. “We will be talking about the digital world, starting from legislation and available technologies, to future technologies such as quantum computers. As this is an event organised under the auspices of the United Nations, we will discuss horizontal topics with a broader perspective. In Europe, we talk about 5G and 6G networks, about increasing bandwidth, while half of the world does not have access to the internet”, says Krzysztof Szubert.

Marek Kolewski

Read an interview with Krzysztof Szubert on page 2