The 5th edition of the Conecta FICTION event featured an enlightening discussion regarding the potential of artificial intelligence in the audiovisual sector. The seminar featured Sami Arpa, CEO of Largo, Alexandra Lebret, producer Eduardo Carneros, Javier Fernández Fernández, and Silvia Werd Elias, director of Marketing at JUMP.

Let’s see how the discussion went.

What is AI?

Javier Fernández Fernández began his reply to this question with “that is a difficult question,” before going on to state that AI is a “set of techniques that allows us to handle data in order to create processes and get new information. In the end, every process in every kind of industry depends on this information.” He went on to conclude that “the main goal at the end is to make the correct decisions.”

So, in essence, artificial intelligence and machine learning represent the use of complex computer data science to extrapolate new information that can help users make more informed decisions.

Where do you find the data for artificial intelligence systems?

Silvia Werd Elias took this question and stated that there is a huge range of sources that can provide invaluable data pools. She stated that her company doesn’t use personal data but rather the demographic data platforms collect about their users.

“We gather all the information on user behavior and user consumption habits because you don’t watch the same movies on a Friday night as on Monday morning, for example”.

This approach is standard for all AI industries. The accuracy of artificial intelligence systems depends on the volume and quality of the data that these systems can analyze. This is why data has been called the ‘new gold’.

How does AI filmmaking work?

Largo.ai CEO Sami Arpa opened by explaining that it is very hard to create a system to analyze the broad range of data sources and types that are involved in the filmmaking process.

“By learning from data, we aim to provide three main areas of insights. First of all, contact insights, to look at your content from different angles.”

He explained that these insights aim to provide filmmakers with added objectivity which allows filmmakers to examine their content from different perspectives. “For example, the system is able to show how comedy evolves throughout the course of the script”. By seeing a visual representation of this, filmmakers will be able to ascertain if the film is fulfilling its goals via a different form of representation than they would otherwise have.

Arpa then continued that the other two main areas are testing, where the AI analyzes all character relations and then provides a list of potential actors suitable for the role, and finally, financial forecast where producers can see gross earning projections and how this will be affected by any changes they have made.

What do the producers think of AI?

Producer Eduardo Carneros was asked how he felt as a producer about AI and whether he was ready to use it as a tool to help in the filmmaking process. He stated that for “almost 5 years, I have been looking for this kind of tool”, continuing that “I wanted to check the screenplays that I was developing before taking them to the marketplace”.

Carneros went on to say that AI is a powerful tool to help improve the accuracy of the filmmaking process. It offers a number of tools that can be used to assess aspects such as the script and to, in some cases, identify areas where scripts can be improved, etc.

These sentiments were mirrored by an overwhelming majority of the European producers that recently undertook a case trial of Largo.ai’s data-driven moviemaking service.

Of the 22 producers and production companies that took part, 60% said that they would continue to use AI while a further 20% stated they might continue using it but just not at the current time.

Given just how far artificial intelligence has to go in terms of developing its full potential, these figures highlight just how far it has already come and just how powerful a tool it is in helping filmmakers to get their films right.

Can AI companies develop systems for streaming services?

This question was put to Sami Arpa who started by stating the difficulty of trying to compete with platforms such as Netflix that are using their own data to develop their own in-house AI systems. Arpa explained that since these platforms are refusing to share their data, this obviously makes things very difficult.

However, he stated that data is available from other sources which allow non-steaming AI platforms such as Largo.ai to train their systems. While the environment is certainly competitive, social media platforms such as Facebook and Instagram are happy to provide their user data for a price.

Meanwhile, marketing companies such as JUMP are able to analyze data from a wide range of sources to extrapolate relevant insights. Silvia Werd Elias points out the difficulty and importance of getting “the right algorithm” in order to make predictions such as which films and TV series will do well over the coming months, etc.

This allows companies such as JUMP to advise broadcasters as to what their users will want to see and how to retain users to stop them from leaving the platform, etc.

Solving the distributor problem

When asked about the power of AI to help with distribution, Arpa introduced a new tool that has been recently added to Largo.ai that helps solve one of the biggest problems facing movie production companies.

“One thing that is very strong with AI is its ability to identify which film is suitable for which distributor. If you add AI in this connection point, with the system matching the most relevant distributor to each film, this will give both parties more confidence that the film has a bigger chance of success”.

To watch the complete debate in full, click this link for Largo at Conecta Fiction.