The “I” of intelligent machines and the possibility of their integration into human societies

September 29 2021

We often wonder whether the rapid development of modern artificial intelligence can lead to the possibility of building “intelligent” machines that will not only simulate human intelligence in individual areas —often far exceeding it— but will have qualities and abilities that will enable them to participate in most activities of our way of life. According to the theory of artificial general intelligence (AGI) or “strong AI,” we would then move from today’s computer systems to android robots that could join our societies in a variety of upgraded roles.

This issue concerns researchers in the field of cognitive science, while it is approached more broadly through science fiction literature and films and is examined in the context of debates that address complex philosophical problems. Predictions are risky, because they presuppose the answers to open questions about whether these machines could eventually acquire real consciousness, as well as the desires and emotions that characterize the members of human communities. Specifically, it is argued that in order to speak about machines with which we will interact as we do with our fellow human beings, it is first of all necessary to attribute to them a form of self-consciousness, expressed by the emergence and manifestation of a sense of self (“I”). Thus, we may later be able to treat them as quasi-persons with a distinct identity, endowed not only with language and rationality, but also with agency and at least limited autonomy, with interests and rights.

Here, the opinions of philosophers are divided. Those who adopt functionalist conceptions of the mind believe that the formation of self-consciousness requires no more than the full development of networks for the reception and processing of stimuli from the environment, to produce appropriate behavioral responses. Others believe that a particular biological constitution, which only naturally evolved organisms possess, is indispensable for having a self, but also for being aware of the unique qualitative features of experience caused by external stimuli and leading to genuine self-conscious reactions. Moreover, those who embrace the theistic belief in the existence of a transcendent Creator rule out the possibility of constructing intelligent systems with the characteristics of human persons.

Now, if we could indeed design intelligent machines with the prospect of developing adequate self-consciousness and some form of personal identity in their bionic bodies, we should be seriously concerned about the roles that will be assigned to them once they are integrated into our societies. The key question is whether they will not be mere servants or subordinate employees and assistants, but whether they could become partners or even companions in various aspects of our lives and eventually attain autonomy or even complete independence from our volitions. Then, they should be treated as having rights, but also clear obligations to human beings, so that we will be protected from the threat of their regarding themselves not only as our equals, but also as would-be masters with the capacity to enslave us. This is why care must be taken, when programming them, to include in their software a system of moral principles and values that will guide them, as well as to institute the appropriate legal framework for their social integration, which we will constantly monitor and adjust, if necessary.

Stelios Virdivakis participated in the discussion entitled "Humanity - AI Symbiosis"