Elizabeth I never sat for a portrait she didn't approve. Every ruff, every jewel, every allegorical prop was a decision about power. A new London exhibition at Philip Mould and Company unpacks the Tudor queen's image-making apparatus, and it reads, uncannily, like a 2026 brand strategy deck.
Sovereign Image Control in the Age of AI Aesthetics
The parallel becomes sharper when you stack it against a 2026 arXiv paper by Johnson et al. evaluating AI-generated images of cultural artifacts using community-informed rubrics. The paper argues that measurement frameworks for AI image quality have systematically excluded marginalized communities from defining what counts as accurate or dignified representation. Elizabeth's court had the same problem in reverse: the queen held absolute veto power over her own likeness while her subjects had none. Both scenarios are about who controls the rubric. And the rubric is never neutral.
From Tudor Courts to AI SEO: The Representation Economy
Then there is The Verge's investigation into AI SEO, where marketing firms are now attempting to influence how large language models describe products and companies. This is portrait-making for the algorithmic age. Instead of commissioning a flattering miniature, brands are seeding training data. The queen's propaganda apparatus has been democratized, then immediately re-monopolized by whoever has the budget to game the model. A 2026 arXiv paper on LLMs as behavioral simulators found that even when model outputs look right, causal effects on perception can go systematically wrong. The image looks sovereign. The power underneath it may not be.