Sora, Synthesia and the synthetic-media step-change
Text-to-video is here, AI presenters are reshaping corporate comms — and a question worth sitting with: how much of the drudgery in creative work is the art itself?
With OpenAI's Sora and Synthesia's AI presenters stealing the spotlight, the big question looms: what will this new creative era for human-made media look like? From massive infrastructure investments to cost efficiencies and regulatory shake-ups, AI is rewriting the playbook.
The money behind the models
US tech firms are set to spend over £300 billion on AI infrastructure in 2025 — Amazon alone plans around $105 billion this year, even as short-term capacity concerns nudged its shares down around 4%. China's DeepSeek dominates earnings calls, with CEOs fielding nonstop questions about its cost-cutting techniques.
Regulation and the semiconductor read-through
The EU's comprehensive, GDPR-like AI Act contrasts heavily with the US's piecemeal, state-by-state approach — and at the federal level a Trump–Musk double-headed horse is splitting opinion faster than Brexit. In Australia, voices are split between caution and opportunity. A WSJ report flagged a roughly 15% drop in tech and power stocks after DeepSeek's launch, as investors rethink long-term chip demand in a world where models get smarter and leaner. Despite the jitters, the broader consensus is that relentless investment in data centres and semiconductor manufacturing will continue.
So what is Sora?
OpenAI's text-to-video model converts written descriptions into fully animated clips, producing complex scenes with multiple characters and fluid motion. The business impact is real: marketers can generate high-impact visuals without expensive shoots, and filmmakers and game developers can prototype scenes in minutes. It also raises questions about authenticity, copyright, and the evolving role of human creativity.
Synthesia's AI presenters, meanwhile, are transforming corporate communications and news broadcasts. These innovations sit at the heart of a synthetic-content step-change where human creativity meets machine efficiency. Convergence, conflict or confusion? Possibly all three. Synthetic media isn't yet replacing creative output so much as augmenting it — absorbing the repetitive heavy lifting and freeing humans for the artistry, storytelling and nuance machines can't replicate.
Which begs a question: how much of the drudgery — the editing, the painstaking crafting, failing and reworking — is the art itself? Arguably a lot of it. I've been using GPT to help draft these articles, and no amount of prompting delivers a final result that compares to one I'd have begun myself. But my goal here is a weekly self-journal, and to that end AI is genuinely helpful for the drafting board and framework.
Does that make these better? Almost certainly not. Does it make me publish on time, every week? Definitely.