Film’s New Normal? Curating the Future of Artificial Intelligence in Film at EFM Horizon 2020
This year we had the honour of co-curating together with EFM’s Manu Guddait, an intensive two-hour session of talks with international founders, experts and policymakers on the future of AI in the creative industries. Cineuropa reported on the day and its key findings, which you can read below, or in full here.
During a session that lasted two hours, European Film Market (EFM) Horizon attempted to determine whether artificial intelligence (AI) is film’s new normal, with five presentations that explored the presence of AI, in its various forms, at different stages of a film’s life cycle.
During her introductory keynote speech, Maja Cappello, head of the European Audiovisual Observatory’s legal department, focused on the real or imaginary legal issues that the use of AI creates in fiction. By using the example of Ludwig van Beethoven’s unfinished “Symphony No 10”, which is now being completed by artificial intelligence, Cappello raised the question of whether the final work would end up in the public domain or whether a machine could hold the copyright. Given that legal systems vary dramatically between regions, in the UK, it would probably be possible for AI to hold on to the rights, whereas that wouldn’t be the case in continental Europe. Also, another possible issue arises in cases where a machine creates original content: who will be the owner of this product? The machine or the people who “educated” it with the necessary data?
A more contemporary problem is the widespread use of deepfake technology, where people are being replaced by digital doppelgangers or other people, and a series of lawsuits are already under way, as the legislation still lacks the proper tools to face up to this recent phenomenon. Finally, Cappello questioned the efficiency of algorithms that suggest or create content for the users of different platforms, as it is not clear how the AI they use actually exploits the data. However, she also mentioned that 30% of the European content on SVoD platforms won’t be implemented by algorithms.
In her presentation, Monica Landers, founder and CEO of StoryFit, a technology company that provides AI analytics for the publishing and entertainment industries, assured the audience that AI is not a threat to storytelling. She explained that AI could understand and analyse a storyline, but wouldn’t be able to create a new one, as things stand currently. The technology can contribute innovative ideas and make suggestions or notes, as long as the words are on the page so that the script can be examined and processed. Apart from these analyses, through StoryFit, it’s also possible to compare scripts not only in terms of genres, but also to find out more details about each character’s personality, taking into account different factors such as agreeability, cooperative spirit, reactivity or the likelihood of making adventurous decisions. Furthermore, it is possible to create a character network, where the end user can see how all of the characters interact with each other and create their own unique “DNA”, which is useful when it comes to explaining what the relationships and connections between them are. Landers emphasised that AI cannot understand humour, as it doesn’t get the punchlines and overlooks what humans consider funny. Furthermore, it cannot generate a storyline, and she used the sci-fi short film Sunspring, which was fully written by a machine, as an example of this.
In a more industry-specific presentation, Tobias Queisser, co-founder and CEO of Cinelytic, a machine-learning-driven platform that provides data on and predictive analytics of a film’s value chain before it has even been made, showcased a demo. He underlined that the film industry is still using outdated tools to invest in the future of projects, from packaging and financing to marketing and reaching out to the audience, and AI, with a deep learning process that spans over ten years, is able to be fairly accurate, in excess of 80%, when making future estimates. Equally impressive was the presentation of Roderick Hodgson, VP of AI at Simon Says, a website that transcribes interviews, recordings and footage straight from audio, using AI. The company’s technology is able to recognise speech, diarise, auto-translate and punctuate, and so far, more than 100,000 projects have been transcribed in more than 100 languages. Through his tool, content creators are saving money and reducing work hours, facilitating a more streamlined form of production.
The final speaker was Jess Fuselier, community outreach, marketing and data specialist at the Sundance Institute, who focused on the tools that smaller players and creatives can use for marketing and distribution. By reviewing various AI platforms, including some that had been presented earlier on, her goal was to help storytellers to flourish without losing any of their creative power while using them. She also reassured those who are sceptical about creativity being endangered by AI.
According to Fuselier, the beneficial aspects of AI lie in predicting marketplace performance, enhancing marketing insights (regarding the audience, for example), building efficiency in the pre-production process, identifying new story arcs, characters or plots, and bringing more unique, bold and diverse voices to the screen. AI could also become a potential threat when the insights are derived from incomplete data sets, when its use compromises artistic integrity or fully replaces the artist, and when there is an inherent bias ingrained in AI-driven technologies.