email print share on Facebook share on Twitter share on reddit pin on Pinterest


Qui a peur de l'intelligence artificielle ?


- BERLIN 2019 : Ce débat organisé dans le cadre d'EFM Horizon a exploré le statut actuel de l'intelligence artificielle et discuté de la menace supposée qu'elle représente

Qui a peur de l'intelligence artificielle ?
(de gauche à droite) Kathleen Schröter de Fraunhofer HHI, le modérateur A.C. Coppens et Sadaf Amouzegar de RivetAI pendant la conférence (© Thomas Bendig)

Cet article est disponible en anglais.

“The Transforming Force of AI: Hype or High Potential?” This was the provocative and highly sceptical question put to guests participating in this year’s first EFM Horizon panel, on Friday 8 February, a day which the innovative hub of the European Film Market (EFM) dedicated to artificial intelligence (AI). Organised in collaboration with the Fraunhofer Institute for Telecommunications, and moderated by AC Coppens, founder of The Creatives Catalysts, the panel hosted a series of professionals who are directly involved in the development of AI for entertainment purposes, and who presented their findings in an attempt to do away with “AI Stigma”.

(L'article continue plus bas - Inf. publicitaire)

Sadaf Amouzegar, a data scientist at RivetAI, focused on the initial results of her pioneering start-up platform. The company is developing AI software tools – or new and improved data-driven tools – in order to automate script highlighting, budgeting, and other practical aspects of film production, allowing those involved in projects to focus on the creative aspect of their work rather than the tedious and repetitive side of filmmaking. She stressed that AI shouldn’t be considered a “threat” because, ultimately, it is a tool which analyses data to solve problems, all the while enhancing the efficiency of film production. She pointed out that we’re not even close to reaching a 100% AI result at this stage and, for that reason, it would be impossible for technology to replace or substitute human creativity. To emphasize this, she referred participants to Progress Bar, a short sci-fi film comedy directed by Peter McCoubrey, which was screened at Sundance and was written jointly by a human and by RivetAI tools - the AI character’s dialogues were generated over a six-month period by a deep learning model, trained off of a datasheet which included scripts and customer service conversations. She also focused on the future of the platform, which they hope will move beyond the film industry in the next three to five years, with added narrative tools and possible new languages.

Thomas Bendig, CEO of the Fraunhofer ICT Group, spoke along similar lines during his presentation. AI is a tool and not a substitute for human activities, and yet the stigma attached to innovative technology persists, with the image of IBM’s Deep Blue defeating Garry Kasparov in chess, over 20 years ago, still firmly anchored in people’s minds. Today, the AI tools available are capable of far more than just playing chess. They can certainly create content, but can the result be classified as art? By current standards, probably not, as data-driven machines are fed with algorithms from the past, which means they are unable to innovate or create things which don’t already exist. He also insisted that the term “AI” has negative connotations and that “data-drive algorithm” is more appropriate given the technology which is currently available and which is a far cry away from the cyborg images overridingly associated with AI.

Finally, Kathleen Schröter, Head of Marketing and Communications at Fraunhofer HHI, a leading research institute in VR/AR and AI technologies, agreed with the other speakers on the usefulness of AI as a tool. It can facilitate the production of sign language for films, for example, and can also help shape the market, offering products that are anything but homogeneous. The many different varieties of AI can also be applied within other industries outside the realm of media, such as healthcare. AI can also be excellent when it comes to enhancing the physical, manipulative side of Virtual Reality, setting the latter aside from passive entertainment mediums.

As the panel discussion drew to a close, when all the speakers were asked to answer the initial question, they were unanimous in their verdict: AI has high potential, and human creativity will only be enhanced if all those elements which can and should be efficiently automated, are efficiently automated.

(Traduit de l'anglais)

Vous avez aimé cet article ? Abonnez-vous à notre newsletter et recevez plus d'articles comme celui-ci, directement dans votre boîte mail.