AI in Film: Tool or Threat? | CBFF2025 Panel Discussion with Film Cymru & Media Cymru

Описание: Is AI revolutionising the film industry or threatening the livelihoods of creatives?

Watch the full panel discussion from the CBFF2025 (Carmarthen Bay Film Festival) where experts explore the critical impact of Artificial Intelligence on the creative industries.

The video is a recording of an AI discussion forum panel at the CBFF2025 (Carmarthen Bay Film Festival), featuring moderator Brett Aggersberg from the University of Wales Trinity St David (UWTSD) and the festival team, alongside panellists Simon Howlett from the festival, Robin Moore from Media Cymru, and Tracy Spottiswoode from Film Cymru.

The panel discusses the impact of AI on creatives and the creative industries, particularly within film.

Key points and viewpoints raised:

Defining AI as a Tool: Tracy Spottiswoode defines AI in filmmaking as essentially a collection of tools used across the entire production chain (development, pre-production, production, post-production, sales, and marketing). She highlights examples like using ChatGPT for generating ideas (with the caveat of incorrect facts) and apps for remote location scouting, but stresses they are not discussing "AI robots".

Balancing Optimism and Cynicism: Robin Moore (Media Cymru) stresses the importance of balancing optimism about AI's opportunities with cynicism about its drawbacks. He views AI as an "engine for changing modality" (e.g., text to speech, image to video), which streamlines production but carries unintended consequences like hallucination, bias, and the degradation of trust.

AI as a Replacement vs. a Tool: Simon Howlett (CBFF) expresses concern that AI is no longer just a tool, but something people are being asked to train to replace them. He argues that AI "recycles," it doesn't create, and should never replace human creativity. He advocates for continually supporting and championing the creative and the artist.
Assisting vs. Replacing (The Core Struggle): The panel discusses the current industry struggle between AI assisting creators and potentially replacing jobs.

Film Cymru's Stance: Tracy Spottiswoode shares Film Cymru's position, which recognises AI's potential for economic growth and productivity but insists it should not come at the expense of those working in the creative industries. They are concerned about text and data mining for training AI models without the copyright holders' permission, citing a problematic bill and a quote from Baroness Kidron: "we don't own it, so we shouldn't be selling it for free".

"Human in the Loop": She references Dr Angus Finn's research paper for Senorio, which emphasises the need for a "human in the loop," ensuring the creative remains at the centre of the process.
Creative Opportunities: Robin Moore acknowledges external pressures from other media (like computer games and "machinima") and younger generations embracing AI for independent filmmaking. He advocates picking battles against tools that act as a "magic button" and take the human out of the loop, instead supporting tools that allow creators to realise a vision and create new creative ambitions or audiences (e.g., the APOT lip-syncing project for dubbing).

Ethical and Environmental Concerns: Tracy Spottiswoode raises the significant environmental concern of AI data centres, which are predicted to consume the same amount of water globally as Denmark in a year.
Democratisation and Trust: The discussion turns to the messy democratisation of AI, noting that many tools are behind a paywall, disadvantageous to students. Robin Moore worries that AI will negatively exaggerate the social change caused by social media, particularly eroding trust in images and media, making it harder to discern quality or truth. However, he also notes that AI democratises content creation, giving access to creative outlets that people, like his son with dysgraphia, didn't have before.

Need for Training and Regulation: The panellists agree that training is essential, focusing on how to adapt existing professional skills to direct AI tools effectively, rather than being replaced by them. They conclude that it is crucial to lobby for regulation and mitigate the worst aspects of AI, preventing it from becoming a "quick way for lazy people".