Robots can mount, move, pack things up, do heavy work, replace human beings for routine or dangerous tasks. But they cannot make up for  the intuition of a journalist, the creative genius of a writer, or the emotional intensity of a good screenplay. Is this still true?

In the media business, Artificial Intelligence (AI) has been widely tested with remarkable results. It’s a robo-journalism software which for instance prepares in a few minutes most of Associated Press takes about corporate earnings and financial news. The editorial staff stated mistakes are far less frequent now, despite the output has increased about tenfold.

Jeff Bezos introduced AI at Washington Post in 2016, using proprietary technology Heliograf to generate hundreds of pieces about Rio Olympics and US Presidential election. The idea is not only to automatize writing, but also set journalists free from the duty of analyzing tons of data and info sources to file top articles.

What if AI could shoot a film? IBM made an interesting trial, leveraging Watson to launch a three-minute trailer to promote the thriller Morgan. The system processed hundreds of horror movies, until the algorithm could learn which scenes, characters and atmospheres better convey fear – and accordingly select the most frightening Morgan frames. The final product perfectly hit the target.

It’s a trailer, so we are still far from having a full artwork from a machine. But, if you are dreaming of writing or publishing a book, there is an algorithm – it was developed in the US by a commissioning editor and a text mining expert – to predict which novels may turn into bestsellers with 80% accuracy. Prior calculation on some J.K. Rowling or James Patterson works predicted success at 90%. If you ask the algorithm about the magic ingredient, it will answer that 30% of bestsellers is built around just one or two topics, nailing readers’ energies and attention.

There are super efficient algorithms to create the perfect first sentence, or overcome creative standstill thanks to machine learning techniques or collaborative networks. Another possible AI application is to contribute to generate sequels for series such as Game of Thrones or Star Wars. It’s a matter of processing thousand of pages and images to find new narrative plots or evolutions.

Although computers are not yet ready to retire journalists, screenwriters or novelists, create a quality storytelling is the next challenge for AI. And the competition between humans and machines might be surprising.