The endorsement of the screenplays will be dealt with by artificial in

[better21]

According to Nadira Azermay, founder of the Belgian company ScriptBook, dealing with artificial intelligence technologies, the same film company Sony Pictures in the period from 2015 to 2017 could save millions of dollars of its funds if it used the computer algorithm developed by them, rather than relying on people engaged in approval of certain movie scenarios.


Before adopting or abandoning a scenario, film companies usually take into account several factors at once. Of course, first of all, the interestingness of the history itself is considered. Estimated the economic potential of the project - will the film be able to recoup the costs of production and collect a large cash register. But another important factor is who is writing the script - is this person a famous screenwriter, behind which is a whole bunch of successful projects.

One way or another, people make the final decision. Sometimes these people make mistakes. Sometimes very large, costing the film companies hundreds of millions of dollars. Movies in which large amounts of money are invested simply do not pay off. Nadira Azermay, in her turn, is sure that in the future these problems can be avoided, since the function of approving screenplays can be assigned to artificial intelligence.


Using the example of the same Sony Pictures, Azmerai notes, developed by her company and trained on the basis of 6,500 script scripts, the ScriptBook algorithm was able to determine 22 of the 32 box-failing movies of Sony, which indicates its very good predicative ability.

"If Sony used our system, it could weed out 22 deliberately failed projects," Azermay said.

"With the help of the artificial intelligence system we want to create a real revolution in the business of storytelling, giving the opportunity to investors, producers, publishers and filmmakers to reduce all risks associated with the production of a film," Azermay adds.

To develop its ScriptBook system in 2016, the company raised 1.4 million dollars of investment. The current version of the algorithm works as follows: a PDF file with a written script is downloaded to the cloud server. The system conducts a detailed analysis of the document in about five minutes, then reports the results. They show the expected rating of MPAA (rating system of the American Film Association), a potential audience of the future picture is predicted, including gender and race, and, most importantly, forecasts are made regarding the potential box office of the future film.

"When we showed the results to customers, the first question that was always asked was: how can it be possible to just give the computer a script and get such a detailed analysis?" Says ScriptWare expert in the theory and methods of data analysis and ScriptWare processes, Mishil Rulens.

It's all about machine learning. The initial instructions are given to the program by the person, but after the machine goes through the process of self-study and creates massive databases on the basis of which it can very quickly conduct the required analysis.

Publishers, too, will be able to find their own benefit from using ScriptBook's algorithm, says Rollens.

"They have to take a big risk, releasing films in different markets. Now the adoption of such decisions is subjective: they just read the script and try to listen to their intuition. But we want to mitigate the risk by adding here objective parameters that will tell publishers more about the real potential of the project. Expertise is always important, but it is even more important to back it up with real indicators. "

[white light]

Perhaps filmgoers could use their own AI algorithms to decide whether or not to go and see these AI script approved movies.