In a world where AI can generate infinite images, videos, and words in seconds, something feels missing. The outputs are technically correct but they’re off. Generic. Soulless. What we’re drowning in isn’t creativity; it’s AI slop: the flood of low-quality, forgettable content that lacks judgment, nuance, or taste.

Enter Taste Labs, a new startup with a bold mission: to end AI slop by teaching AI to have taste. And behind this vision is Thais Castello Branco, a creator who understood that the real frontier of AI isn’t generation, it’s judgment.

Can AI help save creativity instead of replacing it?

Artificial intelligence has sparked one of the biggest cultural debates of our time. Artists worry about their work being scraped without permission, writers question the future of authorship, and filmmakers are watching AI generate images and videos in seconds that once took entire teams to produce.

AI has nailed objective domains. It can solve math problems, write code, and identify objects with precision. But when it comes to subjective questions, what fits? what feels like you? what’s GREAT? AI is still stumbling.

AI makes creation cheap, but judgment remains rare. And when judgment is scarce, slop fills the void. The result? Clashing color schemes, bland corporate writing, and endless streams of content that look right but feel empty. We’re not running into a shortage of ideas, we’re running into a shortage of taste.

The conversation often feels divided into two camps: those who see AI as an existential threat to creativity and those who believe it will democratize creative expression. But what if there was another way? What if artificial intelligence could be designed not to replace artists, but to strengthen the relationship between creators and their audiences? That is the question at the centre of Taste Labs, the AI company founded by entrepreneur Thais Castello Branco, whose vision is less about generating endless content and more about preserving the value of human imagination in the AI era.

A different philosophy for AI: Building the Taste Layer for AI

Much of today’s generative AI is built on scale. Bigger datasets, bigger models, and more content generated at unprecedented speed.

The result has been what many online now call “AI slop”: an endless stream of generic images, videos, and text that can be produced almost infinitely but often lacks originality, context, or artistic intention.

Taste Labs proposes a different model. Rather than training giant models on vast amounts of internet data, the company argues that creators should own the AI systems built around their own work. Instead of scraping intellectual property, AI should operate with permission, collaboration, and creative intent. In other words, the artist remains at the centre.

Tastelabs is building what they call the data and infrastructure layer to give AI models and agents personalized taste.

Here’s how they’re doing it:

1. A trust-propagation model for tastemakers
Instead of scaling annotators from general labor platforms, Tastelabs seeds their network with vetted people who have strong aesthetic judgment, designers, artists, creators, then lets them nominate others. This creates a community of true tastemakers, not just raters.

2. Products for frontier AI labs
They sell preference datasets, reasoning data, rubrics, and evaluation environments to foundation model providers. This helps AI companies build models that understand subjective quality from the ground up.

3. Tools for creative companies
For app-layer companies building AI agents, Tastelabs provides context and verification tools so outputs are more on-brand, more creative, and less generic.

4. Starting with Design
They’re beginning in design — the domain where taste is most visible. Avoiding clashing colors, generic layouts, and soulless visuals is where AI needs judgment first.

A manifesto for judgement

One of the ideas that repeatedly appears throughout Taste’s philosophy is the importance of judgement. As AI makes creation increasingly effortless, the company argues that human value will lie less in producing infinite outputs and more in choosing what deserves to exist.

In this future, taste becomes a creative skill in itself. Curation, perspective, storytelling, and intentionality become more valuable precisely because AI can generate almost anything.

Thais Castello Branco: A creator who sees what’s missing

Thais isn’t just a technologist, she’s a creator who understands what makes work feel human. Her background spans design, AI, and creative direction, giving her a rare ability to see both the technical and emotional sides of the problem. She noticed something most people ignored: AI doesn’t need to generate more. It needs to judge better. In her words (and in Tastelabs’ manifesto), “the cure for AI slop isn’t better generation, it’s better judgment.” Taste, not output, is the real frontier.

This isn’t just a business for Thais. It’s a mission to make AI feel right, not just technically correct, but beautiful, distinctive, and human.

Why this matters for Creatives
For us creative girliess – for every designer, writer, artist, and creator who pours soul into their work, Tastelabs represents something hopeful:

• AI won’t replace taste; it will learn it.
• Creativity isn’t about volume; it’s about judgment.
• The future isn’t more slop, it’s more beauty.

Thais is building infrastructure that ensures AI becomes a tool for elevating creativity, not drowning it in generic content. She’s proving that technology can honor the messy, subjective, human art of knowing what’s good.