By Weaam William

The global AI race has entered a new phase of geopolitical significance, marked by China’s DeepSeek emerging as a formidable challenger to ChatGPT.

I was introduced to DeepSeek about two weeks ago. What sets DeepSeek apart is not just its technical prowess but its cost efficiency—delivering comparable capabilities at a fraction of the price. This reflects a broader truth: as technology matures, it becomes cheaper and more accessible. This shift is inevitable, but it raises urgent questions: Who controls the narrative of AI? Whose values are embedded in its algorithms? And how do we ensure equitable access in a world where AI could deepen existing divides?

As a writer and filmmaker I believe story-telling is a way to share the messy, emotional truths that bind us as humans. However, sometimes of course, I am tempted to see how an AI program would write my words.

Often it takes out the emotions, which somehow makes it sound a lot more reasonable, but less authentic; less human.

AI’s “perfection” comes at a cost; the soul of the story.

My biggest contention with AI is its environmental impact, with a massive and growing carbon footprint, leaving our planet dealing with a climate crisis with a new challenge. Will AI models develop to become more eco-friendly?

At Holocene Films, we support green AI, as well as workflow optimisation, using AI sparingly and strategically, rather than as a default.

Three years ago, I experimented with Largo.ai, a Swiss-developed tool, to generate forecasts for a film project. The results were staggering. My screenplay, rooted in deeply personal narratives 100% organic writing, scored an off-the-charts ROI prediction. This data convinced a private investor to back the project—yet every state-funded entity in South Africa rejected it. The irony? That same screenplay was the highest ROI-rated project pitched at Cannes NEXT 2022. The disconnect reveals a systemic flaw, human gatekeepers, often swayed by nepotism or inertia would cling to outdated evaluation models. Imagine if AI’s cold objectivity could replace these biased systems, prioritising merit over connections. Would this democratise opportunity for overlooked creators? Or would it risk erasing the cultural nuances that human evaluators understand? Perhaps a hybrid model of evaluation could work effectively, combining the best of both worlds.

In 2022 I also co-founded our tech startup, Holocene Films, drawn to blockchain’s transparency and AI’s promise of objectivity. But I quickly learned a hard truth: AI is not neutral. Its outputs are shaped by the data it consumes—data historically dominated by Western, male perspectives. This isn’t just a technical flaw; it’s a cultural erasure and screams for the need for inclusivity with the training of AI models.

This is why AI development must be decentralised. Nations like India, Nigeria, and Brazil are now building homegrown models trained on local languages, histories, and values. China’s DeepSeek, for instance, excels in Mandarin and understands contexts Western models often miss. Yet decentralisation alone isn’t enough. Users must approach AI with skepticism. It's “hallucinations”, confidently stated falsehoods, are not quirks but existential risks. When I use AI for film pitch decks, I triple-check every statistic. When it visualises characters, I scrutinise cultural details. AI is a collaborator, not a curator.

Holocene Films is where blockchain and AI meet human creativity and innovation. At Holocene Films, we’re pioneering a model that marries cutting-edge tech with human-centric storytelling, paired with a future forward approach to environmental sustainability. Our approach includes:

Blockchain: Securing royalties for creators, ensuring transparent funding flows, and IP protection;
AI Agents: Automating tasks like social media, budget forecasting, and even editing rough cuts— freeing creators to focus on emotion and nuance.

We used Runway to prototype a sci-fi character’s appearance, iterating through 50 designs in an hour—a process that once took weeks. Yet the final choice was made by humans, guided by gut feeling and cultural resonance. Similarly, Luma Labs helped us pre-visualise a climactic scene, but the director’s vision—rooted in personal experience—dictated the final cut.

As a story-teller I find that the possibilities are limitless.

The rise of AI mirrors humanity’s best and worst traits: our ingenuity, our biases, our hunger for progress. Tools like DeepSeek and ChatGPT are not just products; they’re cultural artefacts, encoding the values of their creators. As filmmakers, writers, and innovators, our task is to wield these tools without letting them erase what makes us human—the imperfections, the emotions, the stories that only we can tell.

AI should make “humaning” easier, not obsolete. Let it handle the spreadsheet while we craft the soul of the story. Allow blockchain paired with AI to democratise access while we fight to amplify marginalised voices via the African Film DAO. And let us never forget, technology is a means, not an end. The future belongs not to those who automate fastest, but to those who remember why we create in the first place—to connect, to feel, to be alive, to change the planet to a better place fo all, to redistribute wealth and empower talent. Find us on holocenefilms.

* Weaam Williams is a Cape Town based writer, director and filmmaker.

** The views expressed here do not necessarily represent those of Independent Media or IOL.

#deepseek

Explainer: Why is Chinese AI startup DeepSeek shaking up the tech world?

OpenAI chief Altman signs deal with South Korea's Kakao after DeepSeek upset

OpenAI launches deep research tool as competition heats up with DeepSeek

The Dawn of AI Geopolitics