Original article was published by learto_ler on Artificial Intelligence on Medium
About the Cosmonauta, the Last Consciousness Being in the Universe
During the last month, I was making a video game on the occasion of the jam I Can’t Draw. My proposal was one made in Twine and whose options directed you to 4 possible endings; however, more than the narration itself, what I emphasized was an idea, which is the purpose of this article.
I start first with the concept and synthesis of the video game. Later I develop in a superficial way some ideas around the concept, taking as a starting point the ideas of Kant and Nicola Parisi. Finally, I give some conclusions about the relationship between video game and philosophy, and how this combination is appropriate to explore, experiment, and analyze ideas.
The story unfolds in an uncertain future where a subject rises from a long period of lethargy. After a series of decisions, two are presented that will mark the inflection point: going to a space station or going to a planet; both have a different ending.
What is remarkable at the beginning of the story is the loneliness, which stands out even more with an infinitely large universe as opposed to the small ship that the player has. In addition to this, there is a lack of knowledge about the past and the universe, which the Cosmonauta wants to know but cannot know.
Towards a metaphysics of AI
In a previous article, I pointed out the difficulty of conceptualizing intelligence, mainly associated with logical and pragmatic thinking, and focused on what makes it possible. Therefore, for this article, I associate intelligence with the reasoning in a general way, and from its base, I conceptualize it as the way to solve problems by choosing the most adequate alternative. This shortcut, however, has the problem of leaving aside what consciousness is, a characteristic element of the Cosmonauta that, due to the scope and purpose of this article, will not be taken up any longer.
That said, rationality is not considered a merely human quality, since humans, and now machines, solve problems. So we should ask ourselves how machines think. Nicola Parisi, in her text Reprogramming Decisionism (2017), discusses a metaphysics of artificial intelligence (AI) and argues the need to understand how machines think. She considers that such thinking is mainly based on input data, mostly derived from other algorithms and that it focuses on verifying a conclusion; that is, without making synthesis and based on the verification of a given premise. This can be observed, for example, in some AI applications whose output is usually conditioned directly to the input data, such as the Twitter algorithm (Lyons, 2020).
However, what is observed in Parisi’s text is the understanding of the state of the art of AI, and of some AI algorithms, so it is necessary to ask a question towards a more general vision and to understand, from statistics, mathematics, and computational thinking, how machines reason and will reason, independently of the algorithm used.
One of the characteristic elements that are important to analyze and is described in the same way in the Cosmonauta is the recursivity. In one of the endings, one discovers that in fact, it is not the first time she wakes up and that she will do everything again, taking up again every action as if it were a new one, and starting for it from a telos, which is to discover what happened.
This telos, which is to discover what is in the Universe, also starts from the question of what happens to those objects that the machine cannot reason about. At one point in history there is no one who can reason or think the universe because in reality there is no one, but even with the machine reasoning and traveling can we say the same? Considering that the machine starts from some already established concepts of principle, it can be considered that no, even with the machine traveling, there is a universe of which one no longer thinks.
The above points out two particularities of AI that are expressed through the Cosmonauta: the recursion with which the machine is always ready to travel, and repeating everything as if it were a Prometheus. And a second point is that this reasoning of the machine only contemplates some input data, so it is impossible for it to understand beyond what is in its code. So how can a machine know something a priori, regardless of the input data?
This conclusion also points out the need to think and analyze more about what is recursion, but also to think about what elements make them particular, besides the need to study a metaphysics proper to AI and, therefore, to a possible consciousness.
Philosophy and video games
I close this article to start a discussion that will be developed throughout other texts, and it is the advantage that the video game has as a means to analyze and develop abstract concepts such as philosophical ones. This support, which can be supported by text, images, music, and interactivity, can give us an approach to different philosophical and abstract ideas, hence the need and urgency to develop more video games whose purpose is to question and analyze, beyond entertaining, without the latter being relegated, hence the challenge.
You can download Cosmonauta in: https://leartoler.itch.io/cosmonauta