Training Data How Google’s Nano Banana Achieved Breakthrough Character Consistency
32 snips
Nov 11, 2025 
Nicole Brichtova, the product lead for Google's Nano Banana, and Hansa Srinivasan, the engineering lead, delve into the groundbreaking character consistency of their AI image model. They share their journey of creating a platform where users can see themselves in vibrant AI worlds. The duo emphasizes the importance of human evaluation and data quality, reveals unexpected community uses, and discusses the model's whimsical name. They also touch on the future of visual AI, advocating for accessibility and creativity as essential companions in technology.
AI Snips
Chapters
Transcript
Episode notes
Sketchnotes Unlock Family Learning
- A user hacked Nano Banana to create sketchnotes from technical lectures to understand a chemist father's work.
- That workaround enabled conversations between family members about complex research.
Personal Red Carpet Aha Moment
- Nicole tested Nano Banana with a selfie and a red carpet prompt and felt the result looked like her.
- That personal moment convinced the team the model achieved real character consistency.
Prioritize Human Evaluation For Visual Models
- Use human evaluations for subjective visual quality like faces and aesthetics rather than relying only on automated metrics.
- Combine human evals with internal eyeballing and community testing to capture emotional impact.