Hume’s guillotine and intelligent technologies

Human-Intelligent Systems Integration https://doi.org/10.1007/s42454-021-00035-1 RESEARCH ARTICLE Hume’s guillotine and intelligent technologies Pertti Saariluoma1 Received: 14 April 2020 / Accepted: 28 June 2021 © The Author(s) 2021 Abstract Emerging intelligent society shall change the way people are organised around their work and consequently also as a society. One approach to investigating intelligent systems and their social influence is information processing. Intelligence is information processing. However, factual and ethical information are different. Facts concern true vs. false, while ethics is about what should be done. David Hume recognised a fundamental problem in this respect, which is that facts can be used to derive values. His answer was negative, which is critical for developing intelligent ethical technologies. Hume’s problem is not crucial when values can be assigned to technologies, i.e. weak ethical artificial intelligence (AI), but it is hard when we speak of strong ethical AI, which should generate values from facts. However, this paper argues that Hume’s aporia is grounded on a mistaken juxtaposition of emotions and cognition. In the human mind, all experiences are based on the cooperation of emotions and cognitions. Therefore, Hume’s guillotine is not a real obstacle, but it is possible to use stronger forms of ethical AI to develop new ethics for intelligent society. Keywords Hume’s guillotine · Weak and strong ethical AI · Processing ethical information 1 Introduction The new technological reality is at the doorway of human social life. Many qualitatively new kinds of technologies are emerging. Perhaps the most revolutionary aspect of emerging technologies is characterised by intelligence. These new kinds of technologies are able to perform actions that require intelligence from people. Therefore, new technologies can be used to automate or autotomize complex work processes. These new technologies will thus profoundly change people’s work processes and social lives (Ford 2015, Gungel 2012; Fukuda, 2020; Tegmark 2017). Intelligent robots, data-analysing information systems and many rule-based machines can perform tasks that have traditionally been done by people, because they require intelligence (Brigsfjord & Govindarajulu 2020, Dignum 2019, Gungel 2012, Mueller, 2020). In industry, the finance sector, social administration, military, aviation, traffic and medicine, for example, intelligent technologies are able to perform a * Pertti Saariluoma 1 Information Technology, University of Jyväskylä, 40014 Jyväskylä, Finland large share of the necessary tasks (Ford 2014,, Fukuda, 2020; Gungel 2012; Tegmark 2017, Yang et al., 2018). A decade ago, in order to get a new taxation document in Finland one needed to go to the tax office, queue and wait for it for some time. Today it is possible to input a number and get the document back in less than a second. The content and legal form of the process has remained practically the same, but moving the decision-making process from people to intelligent technologies has saved a lot of time. Consequently, tax offices can operate with a much smaller work force. When similar changes are simultaneously emerging all over society, the way people live will change. Technology has a long history of causing social changes. Each new innovation adopted in society has changed the way people reach their goals (Bernal 1969; Headrick 2009). Technological revolutions such as fire, sailing, navigation, cannons, printing, clocks, steam engines, electricity and nuclear energy have led to new forms of work and social organisation (Bernal 1969). Most technology-induced social changes have been incremental or limited, but some have been very extensive. For example, steam energy made it possible to organise industries. Industrialisation led to a period of massive (and ongoing) urbanisation as people moved to cities in search of work. This process triggered changes in social structures 13 Vol.:(0123456789) Human-Intelligent Systems Integration and morals as modern urban life replaced the old nobilitybased social life (Bryson 2020, Dignum 2019, Muller 2020, Leikas et al., 2019). However, some historical social transformations have been problematic. New systems of working and production have necessarily led to new kinds of social structures, though solving the problems of necessary social transformations has not been easy. These changes often take a considerable amount of time, in order to engage in the necessary social restructuring and develop new ways of thinking that are accepted by society. Ideas such as freedom of thinking, speech, marriage and entrepreneurship became the focus of social development and were established in advanced societies, but this was not an easy process. Some innovative people went to the USA and established the foundations of a new kind of society. In Britain the transition was gradual, but in France the outcome was a period of terror (Schön 2013). Moreover, human understanding and adoption of the increased capacities of new technologies has been far from smooth. Military experts and leaders during WWI repeatedly sent troops to storm the killing zones with no other outcome but a huge number of casualties. Obviously, generals at the time did not have a clear understanding of the capacity of machine guns and artillery against storming troops (Horne 1962). History is full of examples of the difficulties of adopting large-scale social innovations. There is no reason to think it would be easier in a combined revolution of ICT and machine intelligence. The only good thing is that societies now have time to think and search for good solutions before they lose their grip on the way things go forward. Since human social actions are organised by ethical rules (which often become law), it makes sense to consider how ethics can guide AI-transformation processes (Bryson 2020; Dignum 2019). People will be replaced by machines; they will lose their jobs and be forced to learn new and difficult things. Different forms of many other future negative attitudes are likely to receive support in the social information sphere. Thus, it is essential to find good ways for transition processes to minimise social problems. A part of the process involves understanding what ethical technologies are like. Can any technology be ethical? Technologies are just electro-physical processes, and it is hard to understand the ethical aspect of the flow of electricity in a wire. 2 Information and representations The core of modern intelligent technologies is their capacity to process information and to let information processes control their operations. Information is information not matter or energy (Wiener 1948, see page 132). Consequently, an important question is: What is information, and how can it serve as a platform for intelligence? The answer to this 13 question could help us understand intelligence. However, the an (...truncated)