Disclaimer: These thoughts are my own and do not represent the views of any company I work for/with, or any clients I work with on a freelance basis.

As a creative, I have my own thoughts and feelings about AI.

I'll start by saying that it does bother me just how much some people are heavily relying on GenAI. Especially creatives: artists, musical artists, filmmakers, animators, graphic designers, writers, VFX artists, etc. But also everyday folk: for family picture edits, event posters, self-promotion, business promotion, to name a few examples.

Whenever I see GenAI creations, I automatically distrust the product, person, and service being offered. Because it's not representative of the thing it's trying to promote. It looks unprofessional, it's cheap, it feels tacky, it's not authentic. It's... fake.

I would much rather prefer a simple photograph or basic poster designed in Word or Canva over something created by GenAI. Because then I know it's an actual representation of what is being offered.

With that said, I am not 100% against AI or GenAI tools. I will admit to using AI-powered tools in pre-production and the early stages of post-production. It can be handy for generating rough ideas, providing starting points when I'm stuck, and helping to organise and sift through dialogue and footage in some cases. However, I never (and will never) include AI-generated content in my final products.

Creatives are rightfully scared that GenAI will replace human creativity. It's obvious to many that GenAI is not at that stage, but if people (creatives, businesses and individuals) continue to rely on it, and accept it, then yes - we will be opening the path for AI to blissfully replace skilled crafts and talents with fictious AI-generated content.

Working with Companies and Individuals that use GenAI

I take great pride in my work, and considering I'm currently producing considerably less work than previous years, I now take a selective approach with the projects I take on. It should come as no surprise considering my approach to GenAI that I will be less inclined to work with individuals and companies that use or rely on GenAI for their output, or openly promote their use of GenAI tools.

My use of Non-Generative AI

My editing software, DaVinci Resolve Studio, has AI tools built-in. These are primarily 'non-generative AI' tools, meaning they are used for content analysis rather than for content generation. These tools are used offline and do not interact with servers or data centres.

• Voice isolation: used for isolating speech from background noise
• Optical-flow / speed warp: used for interpolating frames in other frame-rate footage (e.g. a 23.98fps clip in a 25fps sequence)
• Super Scale: used for up-scaling low-resolution footage to fill the frame (e.g. a 480p clip into a HD or UHD/4K sequence)
• Voice transcription / automatic subtitles: used for rough voice-to-text transcription, never used as-is in final output
• AI tracking and masking: used for applying advanced colour correction, adding animations, or blurring content

I also use other non-generative AI tools, such as waifu2x for image upscaling (which I first used in my uni grad doc film, Fibres), and YouTube's automatic transcription, similar to DaVinci's built-in transcription. I have also experimented with other speech recognition tools like Whisper.

These tools are genuinely useful and serve a great benefit for video production, as I'm able to locate bits of dialogue from the automatic transcripts, or upscale poor-resolution footage to make it fit better with the technical quality of the content. However, I must clarify that I do not completely rely on these tools and everything is still manually reviewed and moderated by myself and/or whoever else is working on the production.

My use of Generative AI

I am conscious of my use of GenAI and try to minimise my use of it. For example, I occasionally use tools like ChatGPT, Google Gemini and Microsoft Copilot to re-organise my ideas on narrative structures for complex video projects, as it can be difficult to build an engaging structure that flows for a documentary or short video piece. That said, these outputs are still heavily moderated and modified to what feels right to me personally.

I try not to fall into the confirmation bias of GenAI, and actually I often go against what GenAI recommends. But it can be helpful in leading me past a road block or suggesting alternative ways or ideas that I (or my collaborators) haven't thought of.

• Do I use GenAI for audio-visual generation? Absolutely not. I refuse to use audio or video clips created by GenAI.
• Do I use GenAI for creating narratives or text? Only for initial stages - I do not believe what GenAI tells me and will always conduct further research, creating my own narrative that is in no way repurposed or paraphrased from the GenAI output.
• Do I use GenAI for organising narratives? Yes, but its output is only taken as initial ideas and not as fact.

There are other areas of my personal life where I do use GenAI a bit more regularly, but I still try to keep my use of it limited.

One such example is for family research. As you may know, I have been researching my family history for many years now (I started at age 11) and the further I've gone back, the more difficult it has become. I have used tools like ChatGPT and Google Gemini to ask questions relating to older records, such as an ancestor who did not have a father listed on his birth certificate, and had his grandfather erroneously listed as his 'father' on a following census. These tools offered the suggestions that he did not have a father, and the grandfather was only listed as such to protect the family secret (and/or it was a mistake by the census enumerator).

I am also learning to build websites and web tools for personal use. I do watch tutorials, have printed reference books, and follow guides online. However GenAI has given me some assistance in debugging issues and recommending improvements. As these are strictly for personal use only, and as a learning opportunity, I do not trust or rely on the GenAI output and complement it with my other training materials. I would of course hire a professional web developer if it wasn't strictly for personal use.

How not to use AI

I had the perfect example of when overreliance on AI goes wrong back in 2025. I had been the one to introduce this person (who I won't name out of respect) to GenAI tools. We had a large amount of interview material to sift through, and we were compiling edit sequences from it. They used ChatGPT to generate a cut-down script, but when we came to compiling it in the edit, ChatGPT had hallucinated and paraphrased the contents of the interview. Nothing matched up between what it offered and what was actually said. The output that wasn't hallucinated was also unusable - with lots of pauses and cuts that made it difficult to flow naturally in the edit.

Editing can be a tedious process and shortcuts can be worth trying to save time. But this was an absolute disaster, and it was a lot more fun to scrub through the footage manually to find what we needed (albeit at the expense of some extra editing time).

Final thoughts

I know that the topic of AI (especially GenAI) is a controversial one. I follow many on social media that are fully against AI and its effects on consumption and the environment. I certainly feel guilty using GenAI, but I am able to use it in moderation and do not rely on it. Some of the people around me do not feel the same way though, and put a large part of their work through it. I do not have that level of trust in GenAI and hopefully never will. It also raises major ethical concerns for the areas in which I work, such as documentary filmmaking, where misrepresentations and fake/generated content threaten the authenticity and genuine belief in the stories and narratives told.

Think about it this way: if you used GenAI content in a serious documentary film, some will question your authenticity as a filmmaker. I would see that as career suicide, and that's why I refuse to entertain that idea. It feels wrong to me personally and against the spirit of my craft.

I personally believe, and hope, that the AI-hype bubble will burst at some point. Prices of hard drives, RAM, SSDs and even SD cards have gone through the roof and made my hobby/side-job even more expensive than it was. 

Hopefully I've explained my position on using (and also not using) these tools.