16 March 2023
On December 29, 2022, the Italian collective Roy Ming published a children’s book titled “La Volpe e il Futuro”. It is a simple illustrated story in which a fox and a little bear build a robot to teach children not to be afraid of technology.
It may seem like an editorial project similar to many others, yet it has attracted much attention and curiosity for one particular reason: the author of the texts, in fact, is the renowned text-generation software Chat GPT, which thanks to artificial intelligence is capable of producing surprisingly coherent and meaningful texts, indistinguishable from those written by a human. Even the illustrations in the book were generated by MidJourney’s Artificial Intelligence, an independent research lab that has developed an AI for creating images based on textual descriptions.
The Roy Ming collective states: “Our goal is to apply innovative technologies such as artificial intelligence to narrative, thereby demonstrating that these technologies can be a useful tool for creating stories for the new generations.”
But what is Generative Artificial Intelligence? And how do Generative Artificial Intelligence systems like ChatGPT and MidJourney work?
We explain it in this article by Social Thingum, who work daily on the development of innovative technologies and artificial intelligence algorithms applied to the business world.
Generative Artificial Intelligence and its Applications
Generative Artificial Intelligence is a branch of computer science that involves generating various types of content. It is a technology based on machine learning, that is, the computer’s ability to accumulate and subsequently process information received from external inputs.
This type of artificial intelligence is based on neural networks, that is, mathematical models that consist of artificial neurons, the basic computational units of these networks. Thanks to these networks, models based on the interconnections of the information received are created, thereby developing recognition databases with which they can identify certain patterns and create similar ones.
The fields in which this technology can be applied are numerous, such as figurative art, writing, drawing, and even music.
The ability to create texts and understand how to communicate a given topic is made possible through NLP, Natural Language Processing, algorithms.
Through these algorithms it is possible to compute the probability of the words in a text and, consequently, predict which word best fits the context of a new sequence.
ChatGPT and MidJourney: What Are They and What Are They Used For?
An example of an NLP application in writing texts is, without a doubt, ChatGPT, used by Roy Ming. The GPT models have been designed with the aim of creating an entire text based on specific suggestions. In this way, from a simple prompt expressed in natural language, the computer is able to create original and well-structured written content.
In particular, ChatGPT is a language model developed by OpenAI which, based on neural networks, studies a text and its meaning. To train this model to analyze texts, human assistance was required to provide the artificial intelligence with prompts that were used as a dataset to set up and improve ChatGPT.
This tool is therefore capable of processing a new input and, thanks to NLP algorithms, generating a coherent output, just like in a genuine conversation between individuals.
Another type of application of Generative Artificial Intelligence is the creation of images or videos. Again, the computer uses a database of images, video sequences, or GIFs to create new content similar to those analyzed. Midjourney is a renowned application tool in this branch of Artificial Intelligence and generates content through a textual or image prompt. In particular, Midjourney is used to create concept art, such as the drawings and illustrations found in the book “La Volpe e il Futuro” or in the children’s book by Ammaar Reshi “Alice and the Sparkle”.
An image from Alice and the Sparkle, generated by MidJourney
Generative artificial intelligence is therefore applicable in various sectors and is continuously evolving. For now, this new technology certainly represents a creative aid in various fields of application and should be interpreted as such.
The development of new generative artificial intelligences is, however, a very complex process that requires continuous optimizations to update and create new databases and neural networks that work increasingly better and that can better calibrate the input-output ratio.
We at Social Thingum, based on our experience in developing artificial intelligence algorithms based on Natural Language Processing and image analysis, know how extremely interesting the work on this innovative technology is.
For this reason, if you want to receive updates on the latest innovations in the fields of Innovation and Artificial Intelligence and understand how algorithms are shaping today’s world, keep following us on LinkedIn.
Martina Pisani, Federico Mandelli
To Learn More
https://www.ai4business.it/intelligenza-artificiale/natural-language-processing-tutto-quello-che-ce-da-sapere/
https://www.fanpage.it/innovazione/tecnologia/la-storia-dei-roy-ming-gli-italiani-che-scrivono-favole-con-lintelligenza-artificiale-e-sorprendente/