No industry is immune to the advancing grasp of AI technology. In fact, what is perhaps more disconcerting than the often borderline creepy nature of AI renderings, is the sheer speed of its progress. AI is having an impact in all areas of moving image, not just creation and concepting but also pre- and post-production.
AI tools can aid creative writing inspiration or write the whole script for you. It can give you the big idea and then give you different story arcs based on that. There are even AI tools being used to help with casting.
It is proving genuinely useful in post-production and visual effects work. There’s an ever-expanding array of tools from: intelligent masking and object removal; digital upscaling of low-resolution images; colour treatments; actor replacement and de-ageing; process automation to sound repair and batch-processing tools.
But perhaps the most exciting (or terrifying) of all is the generative category. It is certainly the most talked about. Where previously we had text to realistic voice generators, now we have text prompts to realistic image generators. Just as ChatGPT is impacting creative writing, virtual assistants and web search, AI platforms such as Dall-E 2, Midjourney, nVidia canvas, Runway and Adobe are advancing at an incredible rate in the field of creative image making.
Purely AI generated photo-real movies have a long way to go – thankfully. At least anything that you might be interested in actually watching. However, there are already some incredible AI developments in moving image such as “Make-a-video” from Meta, Google’s “Imagen Video” and “D-ID” talking avatars. Some of the most disconcerting by far however is Bhuman, which offers realistic AI generated human presenters and automated script versioning. nVidia are pioneering real time photorealistic 3D animation, their latest AI accelerated computing allows for real time raytracing. Not to mention providing the proprietary technology platform for some of the software above.
But what’s all this going to mean for the future of filmmaking and content creation? It’s perhaps not so difficult to identify the limitations to AI. At least by reasonable assumptions based on the power of human ingenuity and originality.
Currently, generative AI learns by ‘scraping’ the internet for all existing content. Whether it’s text, images, sounds, meta data – everything. Entire artists’ back-catalogues are referenced. Fine art, photography, even the work of an entire production company like Pixar is categorised and cross-referenced. If there’s a distinctive style or an original aesthetic, it can be analysed, reproduced and regurgitated as fresh artwork with often truly astonishing results. With the power of descriptive word prompts, in seconds anyone can re-create anything imaginable, natural or otherwise, in the style of anyone. This has very rapidly called into question issues of plagiarism and copyright and there’s some high-profile legal cases being heard right now.
But will AI ever be able to argue as to why one creative concept has the edge over another? Lines of code can engineer the powers of reason and logic… No doubt AI will evolve to deliver autonomous effective communication that answers a brief.
Whether it could ever evolve to create work indistinguishable from true original art, is the big question. Shouldn’t art necessitate a creator if it is to be considered art at all? Will the quality and craft that earns the mark ‘handmade’ become even more relevant and hold more value as ‘humanmade.’
Could AI ever create truly original work that comes from an understanding of the human condition? Could it ever possess empathy or acquire the incredibly complex human emotions of joy, loss, sadness, anger, fear, need? If asked to write a comedy, could it learn what it is to be truly funny? Could it invent believable drama, or plot-twists that make a thriller… well… thrilling? Surely AI would have to possess a level of sentience to even come close and perhaps this raises more fundamental questions in itself.
Even if AI becomes powerful enough to fulfil all of the above, will we humans trust it, value it, or even care? Could it be just as good on the surface? The creative equivalent of lab-grown diamonds?
A future where AI-powered tools are far, far advanced isn’t hard to imagine. We’ll no longer control operations in a physical sense with a mouse or stylus. AI will enable human-to-machine interactions, that although don’t seem far away, are still in the realm of science fiction today. Imagine Blade Runner-style voice commands where you ‘direct’ the AI. Or perhaps for those bravest, ‘neuro-links’ implanted directly to the brain. Perhaps AI assistants will learn the inherent abilities or deficiencies of their users and intelligently adapt.
Just as today’s more affordable cinema cameras and powerful desktop computing power have brought us closer towards a meritocracy for filmmakers… When anyone can direct and anyone can write a script, will more media channels simply be deluged with ever more questionable content than they already are? True visual artists and those with the vision and passion for story, character development, pace and structure will always stand-out from the crowd. Might AI just empower those most talented to stand-out sooner?
In our business it is important to stay abreast of innovation and constantly adapt to new ideas and techniques. Right now, there is a value to time-saving technology and tools that aid in the formation of ideas in all aspects of creativity and design – not just in moving image. New opportunities in communications that don’t yet exist, may be right around the corner.
If the same standard, consistency, and quality of communication can be delivered in less time, or laborious processes can be diminished, then communications companies and their clients will justifiably benefit.
Ultimately though, responsibility, ethics and care are needed. Over time, the more we come to rely on creativity derived from AI, the more human ingenuity will surely be eroded.
One thing is certain… This is just the beginning.