Cats, Robots and Brain Uploads in Stray: On the Trail of Artificial Intelligence

A sleeping orange cat lying on grass. A yellow butterfly is sitting on its head. The players control this cat in Stray.
Calm before the storm in the beginning of Stray.

Introduction

Disclaimer

Stray is a video game released in 2022, developed by “BlueTwelve Studio” and published by “Annapurna Interactive”. In the game, you play a little cat who has been separated from her family and is trying to find her way back to them. Stray has been given a 12+ age rating in Germany.

This essay contains spoilers about the video game Stray.

Welcome to Stray’s Post-Apocalypse

Basically, you explore a post-apocalyptic cyberpunk world in Stray. Bright neon signs and lamps everywhere illuminate the darkness of the underground city, which is shielded from the overworld by a huge dome. Nature has reclaimed large parts of the city. Debris and garbage, such as empty bottles and carelessly discarded packaging material, pile up in every nook and cranny.

Most of the robots that inhabit the city appear conscious and autonomous, going about their daily lives. They work in bars, in clothes stores, as cleaners for the municipal waste company “Neco Corp” or as security personnel. In their free time, they go to clubs, read, make music, knit or reminisce about a time when the people whose “Companions” or servants they were still shaped the cityscape.

But the humans have all long since disappeared and mutated bacteria, tick-like “Zurks”, are spreading within the walls of the city, which have now also taken a liking to metal, which is why they pose a direct threat even to the robots.

On Velvet Paws Through the World of Stray

The player takes control of a small orange cat which, after an old pipe breaks off during an excursion with three other cats and you fall into the depths of the bunker city, finds itself trapped between rust, dirt and robots. The aim is to find a way back to the surface. There are secrets to discover, various riddles and puzzles to solve, couches and carpets to scratch, objects to knock off edges, countless canyons to roam through or a soft cushion to curl up on and take a long nap. In short, to be completely absorbed in the role of cat.

In the course of the adventure, the players learn more about the everyday life of the robots, the danger posed by the “Zurks” and also more about the missing humans, and they get to know different areas of the city. In order to finally leave the city behind and make it back to the surface, you are being tasked by a group of robots, called “Outsiders”, to open the large dome that seals off the city and at the same time literally lead the robots into the light.

The players are assisted by a small flying robot drone called B-12, which helps them to communicate with the other robots and has a few secrets of its own that need to be discovered.

Man in the Machine?

B-12 is encountered shortly after starting the game. As you take your first steps through the city, you will notice glowing messages pointing the way and shaky cameras watching the cat controlled by the players. A short time later, a small drone comes to life by operating a few switches in a laboratory according to the instructions, inserting fuses into the corresponding sockets and thus completing a download.

A strange box-shaped white machine, from which a multitude of cables run, stands in a back room, next to which sits a broken two-legged robot that corresponds to the typical appearance of the other robots in the city.

Brain Uploads to Save Humanity?

Shortly after activation, B-12’s memories are still hazy. Fragments of these memories of an apparent past life can be found all over the city, which ultimately form a coherent overall picture. Little by little, more details and information about itself, the scientist it used to assist, and its own family come to B-12’s mind, and anyone who couldn’t already guess it in the laboratory, or from one of B-12’s stories about itself and its family, is presented with the denouement shortly after escaping the sewers and entering the “Antvillage”. The cat and B-12 find another white automaton there, just like the one in the lab.

B-12 remembers that he himself was once the scientist who wanted to upload himself, or his rather his own consciousness, into a robot as a last resort. However, something seems to have gone wrong during this upload, leaving him stuck until the upload was completed by a small orange cat.

Hints of this twist can be found throughout almost the entire course of the game and, equipped with the knowledge of the “Zurks” and the discovery that B-12 was once a human himself and initially forgot his own past, it raises the question of whether the other robots inhabiting the city might not also be former humans who have uploaded their brains into metallic bodies in order to protect themselves from the plague.

As already mentioned, there is no trace of the humans who, at least according to the robots’ stories, shaped the cityscape and for whom the robots were supposed to serve as the aforementioned “Companions”. The robots themselves indulge in memories of their “ancestors”, as they often refer to humans, miss their presence and often imitate typical human behavior.

Brain Uploads in Fiction

Stray draws on a concept known from science fiction: brain uploading. Similar processes, although the prerequisites or the corresponding processes are sometimes different, can be found, for example, in William Gibson’s book Neuromancer (1984), the film Tron (Steven Lisberger, USA, 1982), individual episodes of the Doctor Who series (for example the episodes “Silence in the Library” and “Forest of the Dead”), in various video games such as the Portal series (Valve, USA, 2007-2011), Mass Effect 3 (Bioware, USA, 2012) and Cyberpunk 2077 (CD Project Red, Poland, 2020) and, to a lesser extent, in the film Matrix (Wachowski siblings, USA, 1999), although the focus here is primarily on a virtual world that exists in parallel and is only a simulation, so that, strictly speaking, no classic brain upload takes place.

This is because a brain upload usually describes a process in which a person’s own physical, worldly body is left behind and, as the name suggests, the brain of a person or being is digitized and uploaded to a computer and henceforth exists in a virtual world on a computer or computer chip or in a robot that can be connected to the net. In some cases, such as in Tron, the body of the person being uploaded also disappears.

The reasons for using a brain upload are manifold. Most often it is either to establish a new existence in a virtual world, especially in scenarios in which the world has become uninhabitable due to wars or climate catastrophes, or the upload serves as a life-saving measure, for example if a person is seriously injured and the “soul” or character is to be saved.

Brain Uploads From a Philosophical Perspective

From a philosophical perspective, this raises the question of what exactly constitutes a person. This raises questions about a person’s personality and identity. What characteristics must a person have in order to be a person? How is a person seen as an individual? Can we still speak of the same person after a change in character?

However, the question to what extent the mind and body are separate entities (dualism) or just two aspects of a common entity (monism) also arises. The so-called mind-body problem is a central object of investigation in the philosophy of mind and explores a variety of questions concerning the interplay between “body” and “mind”.

At its core is the question of how mental states and physical states relate to each other and how they affect each other. Subsequent studies look at the above-mentioned aspects, but also address further issues such as the debate about free will, or the question of whether computers can also possess a mind.

In the course of various theories and debates regarding the technological singularity, especially in connection with emerging and increasingly “intelligent” artificial intelligence, brain upload is seen as a possible method of creating a superintelligence.

Briefly summarized, a basic singularity thesis (in different variations) assumes that due to an intelligence explosion, usually in conjunction with a speed explosion related to computational processes of computers, ever faster and ever more intelligent machines can be created, which will soon be significantly more intelligent than the most intelligent human of all time could ever be.[1]

In this regard, brain uploads are seen as a method that can ensure that the original hurdle of creating a true artificial intelligence with human-like intelligence can be overcome.[2]

This is particularly important with regard to current economic and scientific developments and achievements. The world of Stray does not appear to be particularly far ahead of the real world outside the video game in terms of technology, although brain uploads cannot be carried out in the real world, at least not yet.

In fact, it is highly controversial whether brain uploads will ever really be possible, as this would not only require incredibly powerful computers. Also, despite state-of-the-art technology, it is still very unclear how the brain and its individual parts function and interact with each other and whether we will be able to upload brains at all.

Landgrebe and Smith, for example, argue that the brain is a complex dynamic system that consists of several complex dynamic systems and is itself part of a complex dynamic system and that we cannot model these systems mathematically, which makes emulation of the brain, which would have to precede an upload, impossible.[3]

In the course of Stray, it turns out that, with the exception of B-12, no further brain uploads have been carried out and that the other robots are classic machines equipped with artificial intelligence.[4] The humans were not wiped out by the Zurks, but by an unspecified plague before or while they moved into the underground city, leaving only the robot Companions behind.

In the following, we will therefore discuss the Companions and their real-life counterparts. Of course, there are no robots in the real world that are as advanced as the Companions in Stray, and they are not as strongly integrated into everyday life. However, it is clear which real and current processes and considerations are reflected in Stray.

The Robot Companions in Stray and Their Real-Life Counterparts

Robots in the real world are currently still mainly used in industrial sectors. There are now also a few robots in some service sectors and in the medical and care sectors that vacuum or mow the lawn at home, assist in the care and nursing of sick and elderly people, serve as orientation aids or carry luggage, and in the field of Social Robotics in particular, efforts are increasingly being made to integrate robots that are as autonomous as possible into everyday life.

Nevertheless, all of these examples differ greatly from the robots in Stray. This mainly concerns the features and capabilities of the companions, as these are much more advanced, and less so the appearance and design, although some clear differences can be identified here too.

The biggest external difference is that most of the Companions that can be encountered in Stray have a typically human structure and they also behave like humans in terms of their movement.

By “typically human”, we mean that they have a central body from which a head, two arms and two legs extend and that they are able to move upright on their two legs like humans. Robots that are used for garden maintenance or in industry, for example, do not usually have this appearance.

Although the humanoid design of the companions from Stray also applies fundamentally to social robots[5] in the real world, and at least here the external difference is not so great, human-like locomotion on two legs, on the other hand, is an ability that these real robots have little or none of, as the current state of the art does not (yet) enable a stable form of locomotion on two legs.

Social robots, especially for the mass market, are currently being designed increasingly in such a way that they are either unable to move or are on several wheels and have no movable legs. The rest of their appearance, however, takes on humanoid traits, although it should be noted that there are various degrees of abstraction here and only rarely – mostly in science, but also in individual areas of the industry – are attempts made to build robots that are deceptively similar to humans.[6]

The Uncanny Valley

This is largely due to findings related to a phenomenon first described by Masahiro Mori, which is known as the Uncanny Valley. It describes an uncanny or unreal feeling that is triggered when looking at certain inanimate objects that look like real people or deceased people.

Graphically, this phenomenon is represented by a curve in which a drastic drop in familiarity or affection for the respective objects can be observed from a higher degree of visual correspondence between moving or immobile objects and a person.[7]

Figure 1: Visualization of the Uncanny Valley including the placement of various inanimate entities on the graphs. The various masks refer to Japanese theatre, where they are worn by actors (Mori, 2012, 99).

As Mori notes, this uncanny feeling, and thus the Uncanny Valley‘s decline, is intensified by movement, especially autonomous movement, so that the observer increasingly gets the impression that “something is not right here”. Mori places different objects or entities on the graph depending on the strength of the uncanny feeling triggered.

According to the course and the descriptions, it can also be seen that the affection for the objects or entities increases again after it has exceeded a certain value and finally reaches the supposed maximum – at least in relation to movable things – in the entity of a living, healthy person.[8]

It should be noted that at the time Mori described the phenomenon, there were no reliable empirical studies on this topic. However, more recent studies suggest that the Uncanny Valley affects people’s perception of robots, especially in social contexts, and not only affects conscious judgment, but also social behaviour towards robots.[9]

The Uncanny Valley is often associated with another phenomenon, the anthropomorphization of objects or (living) beings.

Anthropomorphism

This is a human tendency or behaviour in which typically human characteristics and/or abilities are attributed to an object, but also to other living beings and even natural phenomena. The range of attributed characteristics or abilities extends from human behaviour, such as simple movements, to emotions and mental states.[10]

It does not even matter how much the entities resemble a human being and it also does not matter whether we know that it is an object such as a robot, as Gabriella Airenti states:

“[A]ny entity can be anthropomorphized, including artifacts and biological entities such as plants and animals. […] There is no requirement of human-likeness or a high level of complexity. Moreover, the same entity may be treated by the same person alternately in both anthropomorphic and realistic manners, showing that this attitude is independent of the knowledge about the entity that one possesses.”[11]

In everyday use, we are familiar with the phenomenon of anthropomorphism when we talk about “moody weather” or the “laughing sun”, when we accuse the printer of deliberately not printing to annoy us, or when we attribute to animals that they understand us or have supposedly different cognitive abilities.

However, we also see anthropomorphism strongly represented in children who, for example, treat their toys or stuffed animals like real living beings, which is why the tendency to anthropomorphize was long regarded as a childish gimmick or, in the form of animism, was seen as an important step in child development, which is then replaced by causal thinking in adulthood.[12]

Accordingly, the occurrence of anthropomorphism in adults was seen as a cognitive distortion, a categorical error, an obstacle to the progress of knowledge, or a psychological disposition typical of the immature and unenlightened.[13]

However, this view has been increasingly abandoned in recent times and more recent findings suggest that anthropomorphism is a completely natural ability or tendency inherent in all humans, with the help of which we form connections and also explain the world to ourselves, especially from an evolutionary-historical and evolutionary-psychological perspective.

This means that, on the one hand, by attributing mental states to other living beings, we are able to explain or predict their actions under the assumption that they are self-determined agents and that, on the other hand, we try to reduce uncertainty and generally seek meaning in events or actions.[14]

Further, it is also thought that we anthropomorphize because we are predisposed to look for social behaviour even when we know it is not possible for it to exist. This is a kind of inductive reasoning, whereby we infer that similar objects have similar properties and accordingly we see properties in animals, for example, that are similar to our own and infer, for example, that they also feel emotions.[15]

Anthropomorphism also represents a way in which we can establish a connection or enter into a relationship with non-human entities. We do this by addressing and treating these entities as if they were human partners in a communicative context, which automatically leads us to attribute intentionality and social behaviour to the entities.[16]

Robotic Anthropomorphism in Social Robotics and in Stray

Due to the characteristics and abilities that robots bring with them or that are given to them during production, it is not surprising that they play a special role in the anthropomorphization of objects or that we anthropomorphize them particularly strongly and even attribute life to them.

The reason for this is supposedly that, due to their behaviour, their (autonomous) movements and also their appearance, we do not perceive robots as objects, which they have been up to now, but as subjects or as acting beings or even persons.[17]

Triggering Factors

Damiano and Dumouchel appropriately note that the anthropomorphic tendency is triggered primarily by the key factors of human-like appearance and autonomous movement or behaviour, and that it is assumed that a strong realism in either of these areas allows a robot to reach the social threshold and thus be perceived as a social subject.[18]

In short, the more similar an object is or appears to us humans in appearance and/or behaviour, the easier it is for us to anthropomorphize this object.

However, Damiano and Dumouchel also point out the problem of the Uncanny Valley, which means that the effects of the Uncanny Valley on our interaction with robots must be taken into account for the design and development of robots, especially social robots, which are to be integrated into everyday human life to a greater extent than all other robots, so that they do not trigger repulsive or uncanny feelings and can be successfully integrated.

To make this possible, anthropomorphism is increasingly seen in the Social Robotics as an opportunity, or as a fundamental tool to enable a successful human-robot relationship, taking into account the Uncanny Valley. The focus here is on behaviour and movements that are as realistic as possible and less on the human-like appearance.

The reason for this is that the effects of the Uncanny Valley are apparently more likely to be triggered if a robot looks very human-like, but the behaviour and movements do not match the expectations associated with the appearance.[19]

Kate Darling writes:

„When we see something that closely mimics a person or an animal that we’re familiar with, our brains expect it to behave exactly like its counterpart. When it inevitably makes a strange movement or otherwise doesn’t live up to that expectation, we get thrown off. This experience is disorienting, making us feel insecure in our predictions, and could explain the ‘creepy’ factor that some people experience with very humanlike robots.”[20]

In contrast, Damiano and Dumouchel note, anthropomorphization is more strongly triggered by human-like behaviour and movements, even when the robot does not look too much like us. A further advantage of this is that it makes it easier to avoid the Uncanny Valley, as we are more likely to be pleasantly surprised when an object that does not look human behaves like a human.

As can inevitably be seen, the robots in Stray correspond exactly to these specifications or considerations, as we also find them in Social Robotics and also already in parts in real application.

Their appearance is abstracted far enough so that they do not fall into the Uncanny Valley, but at the same time they are clearly humanoid in terms of their behaviour, movements and appearance.

It can also be seen that the robots are more or less advanced and that some robots can perform significantly more complex tasks or actions than others.

Emotional Intelligence in Stray and Real-World Problems

What is particularly exciting is that the companions in Stray also appear to have emotional intelligence and can display these emotions via their face – a display in the head module – and their gestures. It is questionable whether they really feel this way or whether these emotions are not just simulated.

However, allusions in the form of actions, conversations and texts that can be found in Stray‘s world indicate that the robots have evolved and improved, and, at least at the time of Stray‘s story, actually feel emotions.[21]

In the real world, robots are also able to communicate emotions or emotional states via displays or verbal signals, although it can be said here that the real robots do not actually feel the emotions, but that these are predetermined by corresponding programming and are therefore merely simulations. Whether real robots will ever be able to feel real emotions is questionable.

However, it is not surprising that when you look at a robot, you get the impression that it really could be an emotional intelligence or a sentient being. This reveals two major problems in the use of human anthropomorphic tendencies.

On the one hand, this naturally opens the door to manipulation; on the other hand, it seems difficult to distinguish simulated emotions from possible real emotions or sensations, and false expectations may arise. Thus, it could well be that, if the time ever comes when robots can really feel emotions, we won’t be able to detect this at all and may continue to treat the robots as if the emotions were only simulated, which could then lead to morally reprehensible behaviour.

Advocates therefore argue, not entirely without some justification, that in order to ensure that we do not inadvertently violate moral standards, we should already treat certain robots or artificial intelligences as if they already had real emotions or sensations.

On the other hand, we may expect too much from the robots and rely too heavily on their abilities and become disappointed or even endangered because we leave them with tasks that they are not at all capable of mastering adequately. The problem of manipulation is itself twofold.

Firstly, it is of course morally reprehensible if people are deliberately manipulated or deceived. When people’s emotions or expectations are played with and they are led to believe that the robot they are interacting with has supposed feelings, sensations or other typically human characteristics.

For example, if a robot pretends to be in love with a person or to be friends with them and does not understand these concepts at all and is also unable to comply with them, then this is a problem.[22]

But this becomes particularly dangerous when this manipulation or deception has negative consequences for people. Darling says, for example, that one only has to imagine that a robot designed to help children learn language is used for advertising purposes and corporate interests and, for example, incorporates advertising slogans or certain terms into language learning, thus influencing children at an early stage of their development.[23]

Conclusion

The attribution of emotions and the use of anthropomorphism, at least as far as real robots are concerned, must therefore – this much can be said – be viewed with caution and can entail major problems or dangers. In the real world, despite all the technological advances, we are still a long way from robots being truly sentient beings and feeling emotions. Similarly, we are still a long way from uploading brains or talking about artificial intelligence, even if our anthropomorphic tendencies can quickly give us the impression that robots have these abilities and characteristics and are on an equal footing with us.

This is naturally fuelled by science fiction elements, such as the companions in Stray, especially when the story is set in a future that does not seem so far removed from our own present and when the robots seem to be relatively close to the current state of technology compared to other science fiction stories. Players undoubtedly get the impression that the companions in Stray are emotional artificial intelligences and due to the real technical developments, especially in the Social Robotics, the impression quickly arises that the robots in the game differ little or not at all from the robots we can already develop and produce.

On the other hand, it is precisely because of this supposedly near future and the relatively similar technical development that Stray is ideally suited to examining the possible consequences of using such technology and drawing conclusions.

It also raises questions about our own role and the supposedly special, typically human abilities and characteristics.

Finally, it should not be forgotten that Stray is a video game that aims to entertain and be fun, and what could be more entertaining – at least from a techno-philosophical or robot-ethical point of view – than a story that deals with the very questions that science deals with on a daily basis and that makes us think?



Sources & Footnotes

[1] Cf. Chalmers, D. J. (2016). The Singularity: A Philosophical Analysis. In U. Awret (Hrsg.), The Singularity: Could artificial intelligence really out-think us (and would we want it to)? Imprint Academic.

[2] Cf. Chalmers, (2016).

[3] Cf. Landgrebe, J. & Smith, B. (2023). Why Machines Will Never Rule the World: Artificial Intelligence Without Fear. Routledge.

[4] Cf. https://www.thegamer.com/stray-humanity-companions-zurk-lore-story-explained/#how-are-the-companions-connected, last checked: 22.11.2023.

[5] Social robots are robots that are (intended to be) increasingly used in social contexts or relationships and differ from industrial robots, for example. Social robots are designed to interact with humans and should perform service tasks, but also act as interaction partners. For example, social robots can act as shopping assistants in supermarkets, as waiters in restaurants or hotels, but also as conversation partners and companions in care settings or in everyday life. This is in line with the roles and functions of the companions in Stray.

[6] In the sciences, the Japanese roboticist Hiroshi Ishiguro can be cited as a leading figure, who, for instance, built an android based on his own appearance. For more information on this and other androids built by Ishiguro, see https://eng.irl.sys.es.osaka-u.ac.jp/, last checked: 23.10.2023.

[7] Cf. Mori, M. (2012). The Uncanny Valley [From the Field] (K. F. MacDorman & N. Kageki, Übers.). IEEE Robotics Automation Magazine, 19(2), 98-100.

[8] Cf. Mori, 2012, 98-100, and cf. Misselhorn, C. (2009). Empathy with Inanimate Objects and the Uncanny Valley. Minds and Machines, 19(3), 346.

[9] CF. Mathur, M. B., & Reichling, D. B. (2016). Navigating a social world with robot partners: A quantitative cartography of the Uncanny Valley. Cognition, 146.

[10] Cf. Damiano, L., & Dumouchel, P. (2018). Anthropomorphism in Human–Robot Co-evolution. Frontiers in Psychology, 9(468), 2, and cf. Złotowski, J., Proudfoot, D., Yogeeswaran, K., & Bartneck, C. (2015). Anthropomorphism: Opportunities and Challenges in Human–Robot Interaction. International Journal of Social Robotics, 7(3), 347.

[11] Airenti, G. (2018). The Development of Anthropomorphism in Interaction: Intersubjectivity, Imagination, and Theory of Mind. Frontiers in Psychology, 9(2136), 8.

[12] Cf. Airenti, 2018, 2-4.

[13] Cf. Damiano & Dumouchel, 2018, 2.

[14] Cf. Darling, K. (2021). The New Breed: How To Think About Robots. Allen Lane, 94.

[15] Cf. Devlin, K. (2020). Turned On: Science, Sex And Robots. Bloomsbury, 124.

[16] Cf. Airenti, 2018, 11.

[17] Cf. Darling, 2021, 100.

[18] Cf. Damiano & Dumouchel, 2018, 2.

[19] Cf. Damiano & Dumouchel, 2018, 3.

[20] Darling, 2021, 110.

[21] Cf. https://www.thegamer.com/stray-humanity-companions-zurk-lore-story-explained/#how-are-the-companions-connected, last checked: 22.11.2023.

[22] Cf. Sullins, J. P. (2012). Robots, Love, and Sex: The Ethics of Building a Love Machine. IEEE Transactions on Affective Computing, 3(4), 407-408.

[23] Cf. Darling, (2021), 161.

Leave a comment

Your email address will not be published. Required fields are marked *