Taylor Swift, Brain Chips, and a Vogue Supermodel That Isn’t Real

The AI Wars: Now Featuring… Superpowers

Forget Google vs Meta vs OpenAI. The biggest AI war right now is China vs the United States.

Enter DeepSeek, a Chinese AI app that rocketed to the top of the charts in January, wiping billions off American tech stocks in a single day. Faster, cheaper, and spookily effective, it sent shockwaves through Silicon Valley and even rattled Washington. Cue Trump, tariffs, Tim Cook at the White House with metaphorical fruit baskets, and a whole lot of America-first posturing.

Meanwhile, the UK tried to stay in the game with the Alan Turing Institute, only for a whistleblower to accuse it of being toxic and mismanaged. Exactly what you don’t want when the AI Olympics have already started.

Elon Corner (because of course)

Elon’s had… a week.

He accused Apple of favoring OpenAI’s ChatGPT over his own Grok

He lost to OpenAI in a robot chess tournament (yep, the big models all battled it out on the chessboard, and Grok only came second).

Sam Altman, his ex-BFF turned nemesis, launched Merge Labs, a direct Neuralink rival. Brain chips at dawn.

And then, worst of all, Grok’s new “Imagine” tool produced unprompted explicit deepfakes of Taylor Swift. Experts called it “misogyny by design.” Horrifying, and a stark reminder of why guardrails matter.

Listener Question of the Week: Should You Trust ChatGPT with Your Health?

Han wrote in to say they’d been asking ChatGPT for medical advice instead of seeing their GP. James’ verdict: a little rash = fine, heart palpitations = call 999.

We also learned about a poor guy who replaced salt with sodium bromide for three months on AI’s advice and ended up hallucinating in hospital. So yeah, sprinkle a little common sense before asking a chatbot about your rash.

The Guest Spotlight: Seraphine Valora

Our big feature this week was an interview with Valentina Gonzalez and Andrea Petrescu, the brains behind Seraphine Valora, the world’s first AI marketing agency that just made history. Their AI model Vivienne fronted a Guess ad in Vogue – the first time a major brand ran a global campaign with an AI-created model.

Their story is wild: two architecture grads, no funding, late-night uni chats about starting a business. They experimented with AI jewellery campaigns, went viral, got a DM from Guess co-founder Paul Marciano (which they missed for nine months), and suddenly they’re in Vogue and plastered across storefronts in Paris, Milan, Vegas, and beyond.

Some takeaways:

• Their process isn’t just “prompt and pray.” It’s design-led, involves studying a brand’s DNA, staging, lighting, and a whole workflow of AI and traditional photography.

• They overdelivered for Guess, sending thousands of images when asked for a handful.

• Their AI models (Vivienne, Anastasia, Margot…) are available for hire, or they’ll build a custom “digital twin” if you want your exact face cloned into campaigns.

• They’re not trying to replace supermodels. In their words: “Photography is the heart of fashion. We’re here to supplement, not compete.”

• They’re young, clever, and articulate, and they’re sparking the debate: should AI-generated beauty be shaping what we see in fashion? They argue it’s no different from retouched photoshoots. Their critics say it sets impossible standards.

Final Thoughts

This episode was packed: geopolitics, deepfakes, health AI, and the future of fashion. The world of AI isn’t black and white. It’s messy, creative, scary, inspiring, and sometimes ridiculous. Exactly why we love doing this podcast.