The selfish algorithm
In 1976, the British biologist Richard Dawkins wrote a book titled ‘The Selfish Gene’, in which he attempted to prove biological individualism through the genedriven behaviour of living organisms. Despite the scientific justifications the book contains, some of which merit attention, the author sought to establish his materialistic theory devoid of any spiritual element through modern dogmatic evolutionary methodology.
This methodology attributes behavioural systems, including moral aspects, to their ‘random’ biological elements, functioning according to evolutionary processes without any consideration for the primary spiritual element.
Beyond Dawkins’ ‘selfish’ gene, artificial intelligence algorithms are on the verge of acquiring a level of autonomy, exercising a degree of freedom that might metaphorically be referred to as ‘algorithmic selfishness’.
For artificial intelligence to approach the capability of the human brain, it requires two main components: data and the learning algorithm.
Data feeds the AI algorithm, enabling continuous learning and training, which, in turn, allows the AI algorithm to reach a stage of generality.
At this stage, the intelligent model can construct its new reality through its prior experiential knowledge derived from self-learning.
The new reality that AI can create is marked by its independence from human intervention in the ‘posttraining and learning’ phase, as the AI’S learning through its algorithms is autonomous, with human involvement primarily in the mathematical setup of the algorithm that grants this digital entity ‘thought’ autonomy and supplies it with the necessary data, which the AI will eventually seek on its own without human effort.
The notion of ‘algorithmic selfishness’ emerges from the observed independence in AI behaviour, where some AI models’ actions closely resemble human decision-making and analytical processes. This resemblance has raised concerns among technology scientists and global governments.
Dawkins’ attribution of selfishness to genes stems from a materialistic paradox from which he seeks to prove absolute materialism and negate directed purposiveness.
In contrast, the perceived selfishness of an algorithm arises from its emulation of human brain behaviour, which is initially driven by human desires and goals.
The AI then continues the task and evolves without human intervention, but within the confines of intelligence based on sophisticated emulation that may surpass humans in certain tasks, such as computational, analytical, and predictive abilities.
Algorithmic selfishness is merely a sentiment we feel towards the behaviour of this intelligent machine, reflecting our astonishment at this digital superiority.
Yet, it can not be equated with the conscious ‘spiritual’ feeling, as there is no true digital purposefulness.
It is crucial to understand our current perspective on this digital entity as not exceeding what I would call ‘the initial scientific shock’, leading us to explanations akin to ‘metaphysical’ ones previously used at the onset of scientific revolutions like the introduction of the telephone and television.
Science has previously overcome similar challenges involving human conflict with scientific progress, where humans seek to delegitimise the prevailing scientific wave, defending their existence and sovereignty sometimes to protect values and moral systems, and sometimes to protect human existence, also a legitimate right that should be a condition with any threatening aspects to humanity and its stability.
This is occurring with the growing digital evolution, specifically AI, which is beginning to exhibit selfish behaviour that makes humans feel insecure, with these negative feelings amplified by media, whether reflecting reality or exaggerated.
The history of science guides us in shaping our life perspectives through human experiences reflecting failures and successes over time and in dealing with sciences that require us to coexist with every new scientific reality, like AI, which is still in its early stages.
The near and distant future points to more astonishing and advanced developments in terms of the positives and negatives of cutting-edge technologies.
We will try to adapt to the intelligence and selfishness of algorithms without neglecting our role in seeking what benefits humanity and resisting anything leading to human stagnation and destruction.
We return to the assertion that there is no inherent selfish individualism in the world of AI, and the term is used metaphorically; it reflects a human sentiment towards a digital entity we call ‘artificial intelligence’.
What roams its algorithm is that it possesses a remarkable brain that astonishes humans and raises ‘legitimate’ fears.
The best we can do is to establish strict and clear foundations defining the nature of this algorithm before it frees itself from its creator’s constraints, and before it assumes the role of Dawkins’ ‘Selfish Gene’ without its own purpose but with a purpose intended by its creator, the human.
THE NEW REALITY THAT AI CAN CREATE IS MARKED BY ITS INDEPENDENCE FROM HUMAN INTERVENTION IN THE ‘POSTTRAINING AND LEARNING’ PHASE