“The reporter found himself in front of the AI research team after a tip led him to their secret research lab. He was curious about what they were up to and, after a few questions, he was allowed inside. The team was working on a new AI that could learn and evolve on its own. ”
The writer knows how difficult it can be to deal with a blank page. With that moment, in short, when you don’t know how to start, how to start an article, a project, a degree thesis, a document in general. For this reason, when screenshots of short stories on a green background began to circulate on Twitter, many users started looking for the source of those ideas, of those stories that were sometimes centered, other times a little absurd. The mind behind those incipits is an artificial intelligence, the one used by a tool called Narrative Device. It is a project that is as interesting as it is simple. In a minimal interface, the user can enter two words. Starting from these terms, an AI produces the beginning of a story. Only the first few lines, nothing more. Because, as its creator Rodolfo Ocampo explains, “at the base of everything is the desire to create spaces for collaboration between man and machine”.
The collaboration between man and machine
Rodolfo is a PhD student at the University of South Wales in Sydney. He is concerned with artificial intelligence and, above all, with the way in which these systems can stimulate and support creativity. “Imagination – he explains to Italian.Tech – is one of the things that we consider human by definition, one of the last we think away from technology. In fact, there are artificial intelligences capable of being effective in this field as well. To make everything credible, however, collaboration, interaction is necessary “. And, in fact, what the Narrative Device project wants to do is to stimulate the creation of a relationship between man and machine. “There is a continuous exchange – says Rodolfo -. The human being chooses the two terms from which to start, artificial intelligence combines them, perhaps using unpredictable ways. At that point, the ball returns to the human being, who can decide whether to develop that incipit. Apart from the component of fun, of curiosity, there is a relationship that, hopefully, can lead to something, perhaps to a new story “.
Technology: Open AI’s Gpt-3
From a technological point of view, Narrative Device uses Open AI’s Gpt-3 artificial intelligence, one of the most sophisticated AI systems in the world. This model is basically an extremely complex text completion system, trained on everything written on the web, from newspaper articles to Wikipedia. Since December, it has been available to everyone, via API: in other words, anyone can use the technology to build their own applications. And that’s exactly what Rodolfo Ocampo did for Narrative Device.
“I started from Gpt-3 – he explains – but, instead of giving the artificial intelligence a paragraph to complete, as is usually done with these systems, I chose to work on the instructions. Basically, I designed the project through a command: build a short story starting from two terms, those that the user enters into the system. Starting from the input, AI fishes in its boundless database and puts together the most suitable words for the context “.
Is artificial intelligence within everyone’s reach?
Narrative Device is interesting a little because it is fun: try to waste some time and surprise yourself at how creative artificial intelligence can be. Beyond curiosity, however, there is the beginning of a journey. The wide availability of extremely sophisticated artificial intelligence models that are now accessible through software libraries, such as those of Open AI, Google or Amazon, among others, has the potential to bring our relationship with the world to a new stage. ‘TO THE.
“The technology exists, it is available to anyone who is able to use it – specifies Rodolfo -. What is missing, at the moment, are the methods of use, a study of ways to make AI interesting and accessible for users. In the near future the challenge will be more than anything else to build interfaces, to design spaces for interaction between man and machine. What we will see, according to my point of view, is the birth of a series of applications based on artificial intelligence, at the service of users on the most disparate topics. We talk a lot about the metaverse, about the web 3: I believe however that this is the next trend in technology and that it is closer to us than we think “.