Foreword

We are at it again. We do not mean that we are back to xCoAx again, although we are extremely happy that that is also the case. We mean that we, artists, researchers, performers, teachers, students, and, more in general, tech-savvy people are dealing with yet another technological revolution that will allegedly change the world. Yet again, there are clashing narratives about such change: on the one hand, our lives will be made much easier, with more free time and less tedious labor; on the other, we are on the brink of extinction by the machinic hand of a technology that seems to be quickly escaping our control.

This kind of discourse is not new. Actually, it begins to feel a bit trite to those among us with many moons of experience in the field. Still, if everybody we know, independently of age, profession, and interests, starts asking questions like “how does ChatGPT work?” by carefully spelling yet another acronym that entered their newsfeed, we have to face the fact that, at least from the perspective of mass media, this impact is unprecedented.

Interestingly, very few people wonder what that acronym stands for, and it is a pity because it may trigger questions that would reveal the real age of the issues at stake. We are not being ageist here: we do not necessarily want to deal only with brand new problems. However, we would like to take pride in reaffirming that we have been at it for quite some time and, hence, we may have something to say about it.

What does the G in GPT stand for? It stands for “Generative”. Are questions around the concept of “generativity” new? Absolutely not: for instance, our dear friend and immensely valuable supporter Philip Galanter was asking and brilliantly tackling the issue more than twenty years ago already. Why are people wondering now about the impact of a generative tool? Hasn’t the concept already been analyzed in all possible ways?

This is where things get complicated and interesting. We may make a big effort in analyzing and hashing out all aspects, facets, and nuances of a concept. Our analysis may be the most thorough, unabridged, and exhaustive endeavor ever brought to completion. Still, such a task will intrinsically be carried out in a specific place and time. We may delude ourselves in thinking that our musings transcend space and time, but even in the hardest of sciences, with its most general equations, a subversive revision process may be triggered by the results of a newly deployed instrument.

Is this what’s happening? Is ChatGPT to generative systems what the James Webb Space Telescope is to astronomy? This analogy is less bold than you might think. After all, both tools enhance our vision into our past, although on very different time scales: billions of years with the telescope, a couple of decades with the chatbot.

This is where the P of the acronym comes into play. It stands for “Pre-trained” and, indeed, points at the billions of digital documents (not years nor stars) with which the software has been trained to detect and exploit statistically significant correlations among words in existing texts and infer relevant rules for the creation of new ones. Are these texts actually “new”? In the trivial sense of never-written-before sequences of words, they are. In the deeper, generative sense of originality and creativity, the jury is still out; sometimes literally, since machine-based creation of texts and images is more and more often the object of legal debates on plagiarism, intellectual property, and copyright.

The last letter in the acronym does not help us at all in this. T stands for “Transformer”, and even if it may remind some of us of the fancy sentient robots from the old animated TV series or the more recent blockbuster movies, this is a much more pedestrian affair about computational operations that “transform” the input into the output. In other words, as it usually happens with computing machines, it is all about crunching numbers.

“Only when computers originate things should they be believed to have minds” wrote number crunching pioneer Ada Lovelace, to stress the fundamental role that human programmers play in the determination of the outputs of digital machines. Such a position was deemed too restrictive by number crunching visionary Alan Turing, whose dreams of conversational artifacts are becoming, or rather have become reality with ChatGPT.

Two centuries after Lovelace, and one century after Turing we are back at it again. What does it mean to be creative? What does it mean to be original? How do machines and humans relate, compare, and compete in such endeavors? The latest exploits of machine learning are so brilliant that they have sparked the debate again, but rather than writing humans out of the generative equation, the very essence of these techniques and processes, inherently based on huge quantities of human-produced data, seems to be putting humans on center stage again.

The intersection between people and computers is busier than ever, now that everybody is talking about it.

We are back at it. Actually, we never left.

Welcome to the intersection. Welcome to the X.