Selected as one of five finalists for this year’s edition of the Sonae Media Art prize for his work Keystone I, II, III, IV, V, Rudolfo Quintas discusses our data mediated culture and the history, philosophy and applications of Artificial Intelligence within this digital era.

The work serves as a kind of mirror, reflecting not the artificial, but rather notions of language, consciousness and memory – facets not inherent in technology – but humanity.

Keystone I, II, III, IV, V is on display at Museu Nacional de Arte Contemporânea do Chiado until 2 February.

Myles Francis Browne – We in the West have been comparatively slow to embrace AI technologies, what do you cite as the cause for the recent receptiveness to AI? 

Rudolfo Quintas – The “thinking machine” was announced by Simon and Newell more than fifty years ago, with the creation of the Logic Theorist program in 1956. Since then, AI has moved out of the university sphere and out of science fiction, towards the engineering of autonomous decisions applied in society. The public debate must be about what we want or not to automate in society. When there is no discussion, governments have nothing to stop them from making the technologies evolve as much they want. In relation to the automation of procedures for political decisions and classification of citizens, the less democratic, more authoritarian and economically stronger countries are in the front line, such as China. There is clearly a desired relationship of power and control with the automation and classification of citizens. False ideas, such as that AI is less biased, are also propagated. These issues have been talked about by researchers who are dedicated to the ethics of AI. Kate Crawford (Microsoft Research), in her lecture AI and The Rise of Fascism, brought current AI research and its political impact into the public domain.

We must remember that the problem is never technology. But what we do with technology. It’s whether we are using technology to free man or to condemn him. For example, privacy, an important concept in the Western world, has become a real problem because there is no such thing as “blind” information. For example, the case of “The Privacy Project”, an investigative piece of The New York Times, recently published. Misinformation, or manipulation of public opinion, perpetrated by (more or less democratic) governments and large corporations, is the main problem. These issues were already spoken of in the 1980s, with the cyberpunk art movement. This movement can be seen as a first reflection on the impact of technologies on the capitalist system and the danger of their use by governments (surveillance, control, authority, classification). We have to be attentive and develop critical thinking. We are no longer in science fiction.

MFB – Western philosophy too has often regarded the self as being the only form of consciousness, often regarding AI as something insidious. In what ways does the work challenge, or reassert notions of the self and consciousness?

RQ – I can tell you about my first discovery related to Eastern philosophy: when I was a child, I started to practice karate. With this, I acquired a new consciousness about understanding the body: listening to the breath, the heartbeat and understanding that the body is beyond my skin. It is something that begins by giving attention to the interior of the body. I discovered that, in eastern philosophy, the relationship of the body with nature and the intuitive thinking resulting from this relationship is privileged. The consciousness is not in the Self, but develops from the relationship with the environment. From here, more complex understandings are developed, including that the ultimate goal of the individual is to cooperate with the collective. This way of thinking about life makes the model of consciousness in the Eastern world different from that of the West. In this sense, the man-machine relationship was imagined in a different way. The West assumed the imaginary of the good versus evil, extending the history of the separation between body and mind. Hollywood movies are an example of this. With contributions, among others, from the history of Christianity and Cartesian philosophy. In the West, AI occupied a place in the imaginary, an evil one. Perhaps that is why we have developed a more insidious relationship. In the Eastern world, the man-machine relationship was early seen as an idea of cooperation.

In “Keystone I, II, III, IV”, bridges are established between these two concepts: the individual and the collective. A collective memory is created and updated daily. And, through these three different sculptures that are on the floor, where each one writes different ideas. It was important to have more than one sculpture writing sentences in order to demonstrate the divergence between the written ideas. In developing this collective memory, the work tells us that our conscience is shared, from which each one has the right to be critical and to establish their own relationship.

MFB – Within this work, AI acts more so as a medium, seemingly mimicking and reproducing human language, rather than producing language itself. How, if so, do the semantics serve to distinguish between artificial and natural intelligence?

RQ – The discourse on what is natural and artificial can be controversial. In most cases, the difference between the two concepts depends on the time window with which we observe it. When we notice the characteristics of this work, there are “natural” sentences that are read from Twitter, and the “artificial” sentences that are generated by the sculptures. Artificial and natural language feed each other. The sculptures positioned on the floor generate new sentences from a word, indicated by the suspended sculpture. For this purpose, the AI system needed to learn the relationship between words, and the relationship between their signifiers, in order to speak based on a context. It was decided that, sometimes, these sentences were more realistic, paraphrasing the context of Twitter sentences with a lot of realism. Other times, they presented a greater degree of creative autonomy. In these more autonomous sentences, moments similar to creative writing transpose what would be expectable among semantic signifiers. They write things that are sometimes absurd or make us laugh, considering the relationships established.

The question I ask myself is this: when sentences are badly written and have little meaning, are we faced with a poorly programmed work, poorly conducted, or are we faced with a society that constantly writes things with little meaning? Or, on the contrary, when these sentences are more curious and creative, is it because the artists and their programmers have done a good job? Is the piece responsible for the meaning the sentences generate, or is it the society from where it learned? With Kesytones, we do not operate in the field of literature. We do not operate in the creation of the meaning of what is written. We are creating a device that leads us to question the relationships between collective memory and individual action. We can, among other things, think about the meaning of so many written messages that have little meaning. With so little clarification about life. If written language was an artifice developed to better communicate, to relate with greater understanding, and to better express our individual characteristics, what is the truth of thousands of sentences with so little critical sense that are generated daily?

MFB – It is interesting that the work publishes its oftentimes nonsensical conceptions of the information to twitter. The platform seems a natural fit given our social media cultural of “Fake News”. Did the work develop this tendency naturally or was it predisposed to producing such comments?

RQ – Both. During the three-month period, the system collected, analysed and assessed the feeling of more than 100.000 tweets in Portuguese society. The work developed a tendency towards “fake news”, because we yet again trained the language model from the database generated from the collected tweets. That is how the artwork learned to write ideas and sentences based on the way people write on Twitter: short messages, many of them without semantic concerns and with an unclear syntax. An AI model performs best when trained in a specific context, a specific subject. In our project, we are not working with a specific subject. Every day, the suspended Keystone generates a representative set of the most frequent words written in the last 24 hours in Portugal. Every thirty seconds, the artwork chooses one of these words and verbalize it. The sculptures positioned on the floor “hear” that word and write their own ideas from what they have learned about that word. We don’t know what the words of tomorrow will be and it is likely that the AI will write things with more meaning if they are words more referenced in long-term memory.

In sentences built on relationships of more distant semantic signifiers, curious ideas occur, that we would have more difficulty imagining or thinking about. Here is one of the sentences: “Christmas is killing toilets”. The sentence is so absurd, that it is curious to think about it.

MFB – Though the processes of collecting and data, that is tweets, and its subsequent analysis reproduction all seems very technological, it also seems to speak to the very human experience that is memory. How does the notion memory find itself in the work and how is it reflective of own capacities to remember?

RQ – Memory is one of the most complex cognitive functions that nature has ever produced. The human being struggles quite a lot when it comes to dealing with memory and this made him create artifices to remember. These artifices are different depending on whether we are dealing with biological memory (cellular memory, motor system, spatial orientation, etc.) and affective memory, where positive or negative sensations are registered as a result of what we experience throughout life. These indicate the way forward. And then we enter the collective level of memory. This memory can be and has often been manipulated throughout history. This is the most fragile and controversial level of memory for the human being. After all, throughout history, the human being has been hiding, altering and manipulating the collective memory. And, apparently, we are terrible at dealing with this kind of memory. People quickly forget the memory that addresses the political and the collective. And this is the level of memory that interested me as a starting point for the creation of this work.

The piece is a non-human entity, which develops a collective memory through human experience.  To create a memory that does not belong to a specific individual, that is neither my memory nor yours, but is a memory of the collective. Depending on how this memory is accessed, read, viewed, different readings are born. At the end of the exhibition, we can access this memory and know the words with the greatest presence in Portuguese society. We can do the opposite reasoning: discover the emotionally less frequent words in Portuguese society and think about why this happens. These are the relationships between art, technology and science that interest me from an artistic point of view. The way in which artistic thought and humanities can use scientific data, which will enable us to read more about human behaviour, so that we can understand each other a little better. This dimension of mirror, of portrait, focused on the human behaviour.