Art in Brief Authenticity and Algorithms: An Interview with Carina Popovici
9 snips
Nov 15, 2025 
Carina Popovici, a physicist-turned-entrepreneur and founder of Art Recognition, shares her journey into the world of art authentication using AI. She reveals how her team trains algorithms to recognize an artist’s style and detect forgeries, drawing on her unique background. Carina discusses ethical concerns in the art market, addressing biases and the role of human experts alongside AI. She emphasizes the need for transparency and responsible frameworks, aiming to foster trust in the intersection of art and technology.
AI Snips
Chapters
Transcript
Episode notes
Misapplied AI Sparked A Public Backlash
- Carina recounts the 2023 De Bresci Madonna episode where one AI (face-recognition) wrongly suggested a Raphael attribution and her model contradicted it.
- The story shows how misapplied AI and media hype can harm trust in AI authentication.
AI Gives Traceable Results, Not Opinions
- Carina emphasizes AI outputs differ from expert opinions because models are scientific and traceable to data and training choices.
- You can investigate dataset and model differences to explain conflicting AI results.
Not All Art Suits Image-Only Authentication
- Some artists and techniques are harder for image-based AI, like dripping techniques or works from workshop environments.
- Art Recognition excludes or treats workshop pieces and non-hand-centric techniques carefully in training.