The artificial intelligence is increasingly more important in our lives, and different industries are investing in this technology to predict what products will get greater benefits.
Perhaps one might expect that these practices would not have as much weight in industries whose success depends on more abstract factors, such as creativity and imagination, but not even art creation gets rid of the influence of algorithms.
In an exclusive, The Hollywood Reporter has revealed that Warner Bros has signed a contract with Cinelytic, the creator of an AI specialized in the management of film projects, and whose main job is to calculate which films will be successful at the box office.
Some may think, somewhat innocently, that the only elements that are taken into account when deciding which films are produced are the enjoyment of the audience, cultural value, and other material aspects.
But at the moment of truth, for companies, the economic benefit is one of the most relevant factors, and the film industry is no exception. Every time a study arises if the film they have been proposed is viable, they consider how much money they can raise at the box office, and some projects do not go ahead because they do not anticipate sufficient profits.
For this, the people in charge of deciding which films are given the green light must take into account very diverse factors, such as the fame of the director and the protagonists, the interest of the audience for such plots, or the popularity of the content existing related to that movie (something very relevant in book adaptations and in the film sagas).
The number of variables to be taken into account for decisions of this type is large and implies hours of work screening figures and statistics. What Cinelytic proposes with its AI is to perform this tedious task in an automated way, making it an algorithm that collects and values all these factors, accelerating the process.
This implies that, in practice, an AI would be part of the decision-making regarding the films that Warner Bros brings to the cinemas. And as expected, controversy has aroused.
When media like Discussing Film shared the news, panic spread among Internet users. In response to the tweet of that medium, some users declared that this decision marks the death of creativity in the cinema, and others said that this confirms that the only art whose existence is allowed is the art that generates benefits.
Warner Bros has signed a deal for a AI-driven film management system which will help decision-making for greenlighting certain films. The AI system can assess an actor’s value in any territory and how much a film is expected to earn in theaters. (Source: https://t.co/4wWwyh66fT) pic.twitter.com/sJZbWkivJI
— DiscussingFilm (@DiscussingFilm) January 8, 2020
But is it really such a terrible decision? Again, it is perhaps somewhat innocent to assume that economic benefit is not one of the main reasons why films are accepted or rejected today (ergo an AI will not change anything in this regard ).
And something that is probably not being taken into account is that the task that this AI would perform is a task that is already being carried out, even if it is done by people instead of algorithms.
The Cinelytic AI is first and foremost a data processing tool. It allows us to simulate different scenarios and assess which guarantees better results: does the audience prefer Anne Hathaway as the protagonist, or would Natalie Portman have more success in this film? How popular is the material on which the film would be based? Will there be more interest if directed by Taika Waititi? At what time of the year are there more possibilities for people to go to the cinema?
Taking into account factors such as these, a study can assess which projects will work best and what alterations they can make so that the success of a film they want to develop is greater. And this does not imply that the AI makes any decision. Simply, the algorithm would provide the data that so far the study has to compile and study manually.
Positive data have already been obtained in the use of algorithms of this type. Other similar AIs such as ScriptBook predicted which films would benefit with 86% success, compared to 44% industry success.
Tobias Queisser, the founder of Cinelytic, has pointed out in his statements to The Hollywood Reporter that AI is designed to help with the process, not to make creative decisions of any kind, something that has also been confirmed by Candice McDonough, head of communications at Warner Bros.
And some professionals in the film industry such as Zack Stentz, co-writer of Thor and X-Men: First Generation, have also pointed out that the panic that has emerged as a result of the news is meaningless. As Stentz points out, the use of statistical data in decision-making is nothing new in Hollywood, and despite this, projects are carried out that analytics do not value positively based on the intuition and personal opinion of the people in charge of deciding
Yes, what the article describes is a refinement of analytics that studios already use. My point is that these analytics that purport to tell you which actor is worth how much in these territories are useless compared to the casting intuitions that end up creating magic onscreen.
— Zack Stentz (@MuseZack) January 8, 2020
Internet panic aside, it is true that the use of AI presents certain issues that film studios will have to face. Artificial intelligence makes assessments based on existing information and tends to make fairly conservative decisions based on it. In cinema, this translates into considering which actors, directors, and films have been successful in the past, which does not have to reflect the current preferences of the audience.
It can also lead to major problems. Because AI values new data based on what it has learned to consider positive, the data and proposals of the algorithms can be discriminatory. In fact, it is something that has already happened in various fields: in 2015 Amazon had to cancel a project in which an AI was used to screen resumes because it continuously discriminated against women, and a study conducted by the Government The US had detected racist trends in numerous facial recognition algorithms that use artificial intelligence.
In the cinema, this could have very varied effects. For example, if an AI has learned that most successful films of a decade have been directed by white men, it is possible that it values directors with these characteristics better than people belonging to various minorities, regardless of the predominance of Directors of the first type may be due to discriminatory reasons.