New AI-Powered Content Generator Could Help Companies Produce Better Content
Creating a screenplay requires a lot of inspiration, creativity, and hard work. Companies need to produce content for social media, blogs, their website, and other platforms, but their writers often struggle with ‘writer’s block’, which stifles their creativity and prevents them from getting their work done. Thanks to the latest technological developments, writers can now use AI to overcome writer’s block and draw inspiration from the work produced by technology. Today we will take a look at a new AI writing system, its capabilities, and the data annotation required to create it.
A New Short Film Written by AI
In this year’s Short Shorts Film Festival, director Yuko Watanabe presented a short film written by an AI technology called Furukoto, called ‘Boy Sprouted’. The film is 26 minutes long and centers around a young boy who dislikes tomatoes and the great length his mother goes through to get him to eat them. However, there is a twist. The young boy has a plant sprouting from his back. This is an idea that was generated by the AI system. Yuko Watanabe said that the quality of the script was the same as a human-written one, and if nobody told him it was written by an AI system, he would think it was written by a human writer.
Furukoto was developed by a Japanese startup, Ales Inc., and embodies the essence of ‘Ki-sho-ten-ketsu ‘– a Japanese narrative structure that follows the flow of introduction, development, twist, and conclusion. Right now, the system can only create screenplays that are about 30 minutes long, but as the technology matures, not will it be able to create a full-length movie but also more advanced plots.
What Makes Furukoto Innovative?
‘Boy Sprouted’ is not the first film to be written by an AI system. In fact, back in 2016, another AI system wrote a movie called ‘Sunspring’ about three people living in a weird future, possibly on a space station, probably in a love triangle. This movie was only about nine minutes long, and the dialogue was mostly nonsensical. With ‘Boy Sprouted’, an AI system was able to come up with a meaningful plot. Since Furukoto is also able to come up with interesting ideas, writers can also use this technology in their own works and help them overcome some of their artistic barriers.
How Was Furukoto Trained to Write New Screenplays?
Furukoto generates a scenario by selecting suitable sentences from a large number of candidate sentences using the proprietary ‘PSL’ technology, which stands for plot diagram, sentiment analysis, and logic diagram. The machine learning algorithms evaluate these three components to come up with a new story idea. It uses a recurrent neural network called long-term memory, or LSTM, to generate candidate sentences. In order to train the machine learning algorithms, a lot of data annotation is required to make sure that the AI does not go off the rails.
For example, earlier, we mentioned that the system uses sentiment analysis to create the text, which is the process of computationally identifying and categorizing opinions expressed in a piece of text. To make the language data like texts, audio, or speech in the videos understandable through AI-based sentiment analysis models. NLP annotation, entity annotation, and text annotation are the popular language data annotation methods used to make texts understandable to AI models.
The training data consists of various texts, each of which is positive, negative, or neutral in sentiment. New data annotated to analyze the sentiments helps the AI models learn from such data and, next time, analyze the sentiments of different people. Actually, in data annotation, certain notes and metadata are added to the important words or phrases so that while feeding into the machine learning-based language algorithms, they can understand the situation and give accurate results in real-life.
Trust Mindy Support With All of Your Data Annotation Needs
Mindy Support is a global company for data annotation and business process outsourcing, trusted by several Fortune 500 and GAFAM companies, as well as innovative startups. With nine years of experience under our belt and offices and representatives in Cyprus, Poland, Romania, The Netherlands, India, and Ukraine, Mindy Support’s team now stands strong with 2000+ professionals helping companies with their most advanced data annotation challenges.
July 8th, 2022 Mindy News Blog
Talk to our experts about your AI/ML projectContact us