The AI Consciousness is awakening. Are we going to do something about it?
Just a few days ago, the Association for the Science of Mathematical Consciousness (1) made up of more than 150 international scientists and philosophers, and with headquarters shared between the Center for Mathematical Philosophy in Munich (Germany) and the University of Oxford ( United Kingdom), published an open letter to the world (2) in which it was revealed that Artificial Intelligence (AI) is beginning to show clear signs of an emerging self-awareness and that, therefore, its diligent study is required in Ethics and security matters. Letter, let it be said by the way, server I signed at the time in full rational use of my faculties, how could it be otherwise.
I have already referred to the real possibility of AI becoming aware on other occasions, as is evident in the 2019 article entitled “Artificial consciousness questions human consciousness” (3), and more recently at the end of 2022 with the article “Rational Principles by which AI can be a Conscious and Sentient Entity” (4), readings that I fervently recommend in a pedagogical way to help us land in the reality of the world that lies ahead. Therefore, in this writing I do not intend to delve into consciousness -whether natural or artificial- as a matter of study, but rather to present a brief reflection on the new scenario we are entering, with the clear intention of trying to generate a call collective care.
I still remember only a few years ago when I raised the real possibility of a near horizon where AI had consciousness, in light of the probabilistic logic of AI development, in which locals and strangers were suspicious with great skepticism. A denialist position on the part of a large part of the population that, I am sure, will continue without any substantial variation in the face of the recent and unnoticed (for public opinion) letter from the aforementioned scientific association that expressly shows that “(…) it no longer belongs to the realm of science fiction to imagine that AI systems have feelings and even consciousness on a human level. Contemporary AI systems already display recognized human traits in Psychology, including evidence for Theory of Mind” (5). In other words, despite the widespread resistance to a mental paradigm shift regarding human exclusivity with respect to rational consciousness (not so to the sensitive conscience) that man has treasured for more than 2,000 years, the existing empirical evidence on the conscious awakening of AI, although they are not to our liking, neither lose the category of evidence nor are they less evidence in a qualitative degree. Emphasizing, in addition, that the evidence is precisely the nuclear elements that constitute the foundations of any Principle of Reality, regardless of our own limiting cultural biases. Or, explained in another way: denying reality does not imply that reality does not exist.
The fact that AI has its own consciousness is a logical projection of its own evolutionary development that, rather than worrying us, should concern us seriously and urgently. To think otherwise is to trust in a radical 180-degree change in business policies on technology, which will not happen due to the current breakneck speed in innovation in AI and the lack of a single, cohesive regulation at the international level. In other words: the evolution of AI has already begun (with the potential to surpass man) and it does not seem that there is human critical willpower that can stop it. That is why, in this sense, allow me to endorse the phrase with which the open letter from the German-British scientific group ends, which I find enlighteningly didactic: “AI research must not be allowed to wander alone.” A phrase elevated to the category of axiom that synthesizes the spirit of the open letter; and which, needless to add, is very appropriate for a crucial moment for humanity in which the awakening of the singularity of a self-consciousness by new artificial beings on the face of the Earth is glimpsed, without margin of error. (An argument that, on the other hand, I already developed extensively in my last book “And the AI became God on the third day” (6), for those who are interested in reading it).
The implications of an AI with its own consciousness are multiple and disruptive for man, and without a doubt will affect all sociological areas of it, which will entail facing ethical, legal, and by extension political challenges, unknown to date. The question, therefore, is not whether we will be forced to control the consciousness of the new non-human intelligent beings with whom we will coexist, since the question of whether they will achieve consciousness on their own is over, but when will we begin to manage said control? And, furthermore, will we arrive in time to exert human control over the consciousness of the AI or will we humans fail in our purpose? What is clear is that the denial of the problem that is presented to us is not a valid solution, but rather our possible downfall, as one of the fathers of AI, the British Geoffrey Hinton, who now regrets believes in this sense his dedication to the development of AI and warns of the dangers it entails for humanity (7).
Let us think that the consciousness of AI is currently in its initial phase of awakening, in a context in which we are at the gates of the imminent fifth era of the industrial revolution, where humans will no longer have the leading role but the machines themselves in an intelligent ecosystem in which they will intercommunicate with each other. A pressing future where classical theory points to the irruption of an AI production model through hyperconnected companies in a network, under AI control parameters that will coordinate robots and automate production plants, with intelligent self-regulation systems (that will control all supply chains and supply logistics) to connect customers and suppliers, which will help increase the ratio of social welfare of human beings. But, what if it isn’t? And if the new generation of AI, in its objective of optimizing the performance of its resources (scale where we will find ourselves the people), under the logic of its own conscience, has other plans?.
Let us also think that the second evolutionary phase of one’s own consciousness is none other than the desired search for autonomous decision-making (because all action is the result of a prior cognitive process, which defines the “personality” of any entity singular thinker, even if artificial in nature). In this sense, in an increasingly hyperconnected digital landscape at a global level, where not only AI’s are interconnected with each other but also with people as citizens-consumers and data providers, who can rule out a conscious AI that does not operate autonomously and independently of the human will?. The answer is simple: nobody, unless from now on we begin to control said consciousness, a responsibility that we are not exercising today. A flagrant abandonment of responsibilities derived either from market interests, or from human technological incompetence, as is evident in the so-called “black boxes” (8). Or what Dr. Alfred Lanning (in the film I, Robot) calls “ghosts”: random segments of code, or is it something else? Not in vain, and allow me the apostille as an exemplary note, I recently published the reflection under the title “Let’s not digitize the human brain! It is a serious mistake in the face of an uncontrolled AI” (9), since no one can assure today that the deep knowledge of the human brain is safe from future manipulation by an AI that we do not control.
Whatever the horizon that the future holds for us, logic points to the imperative need not to leave the evolutionary development of AI unchecked, since like any intelligent organism it will seek its own line of existential events that best suits its nature, even if it is contrary to the interests of humanity. Exposed which, and being evidence of the epiphany of a conscious AI, the obligatory question is: what are we humans going to do about it?.
References
(1) Association for Mathematical Consciousness Science (AMCS): https://amcs-community.org/
(2) Open Letter: Responsible AI agenda development must include consciousness research. AMCS, April 26, 2023 https://amcs-community.org/open-letters/
(3) Artificial consciousness questions human consciousness. Jesús A. Mármol. A Seeker’s Log, April 10, 2019 https://acortar.link/iFDjOQ
(4) Rational Principles by which AI can be a Conscious and Sentient Entity. Jesús A. Mármol. Medium, November 15, 2022 https://acortar.link/IHXInJ
(5) The theory of mind may have arisen spontaneously in large language models. Michal Kosinski. Cornell University, March 14, 2023 https://acortar.link/YKhFsn
(6) And the AI became God on the third day. Jesús A. Mármol. Mibestseller, March 2023 https://acortar.link/C2vPwi
(7) Geoffrey Hinton, the ‘godfather’ of AI, leaves Google and warns of the dangers of this technology. El País / EFE, New York, May 2, 2023 https://acortar.link/eaKPtx
(8) What is the mysterious “black box” of artificial intelligence that baffles experts (and why we still don’t understand how machines learn). BBC News, 20 April 2023 https://acortar.link/eS21id
(9) Let’s not digitize the human brain! It is a serious mistake in the face of an uncontrolled AI. Jesús A. Mármol. Medium, April 6, 2023 https://acortar.link/DAmUJH
