Google AI brings The Wizard of Oz into the 21st century at the Las Vegas Sphere

Hollywood Meets Silicon Valley in an AI-Enhanced Classic

When Google AI and Warner Bros Discovery announced their collaboration to reimagine The Wizard of Oz for the Las Vegas Sphere, it marked one of the most ambitious fusions of artificial intelligence and cinema to date. The project, involving more than 2,000 people over two years, transforms the beloved 1939 film into an immersive spectacle that places audiences directly inside Dorothy's journey.

The results are staggering. More than 90% of the original film has been enhanced using AI techniques including super resolution, outpainting, and performance generation. Visitors don't simply watch The Wizard of Oz, they experience it through a 16K LED screen spanning 160,000 square feet and towering 22 storeys high.

When Tornadoes Come Alive

The tornado sequence exemplifies the production's technical ambitions. As Dorothy's farmhouse spins violently onscreen, audiences are surrounded by the Sphere's massive display whilst 560-kilowatt wind machines scatter debris throughout the venue. The sensation is visceral: the house itself feels as though it might lift off its foundations.

"When the request came to us, I was almost jumping up and down. This is the best opportunity to showcase the magic that we develop using AI," says Dr. Irfan Essa, principal research scientist at Google DeepMind and director of its Atlanta lab.

It's a far cry from MGM's original 1939 production, yet every enhancement serves to preserve rather than replace the film's authenticity. The project demonstrates how AI can expand creative possibilities whilst respecting artistic legacy, much like Google's broader AI initiatives are reshaping user experiences.

Preserving Dorothy Without Rewriting History

Ben Grossmann, who oversaw visual effects, understood the stakes perfectly. His approach was surgical: AI served as a last resort, deployed only when traditional CGI couldn't maintain performance integrity. The results speak for themselves. Judy Garland's freckles, once softened by celluloid grain, now appear in ultra-high-definition. The Scarecrow's burlap weave looks almost tangible.

"This was a huge and complex task. In some scenes, the Scarecrow's nose is barely 10 pixels, that creates major technical hurdles," explains Steven Hickson, director of AI foundation research at Google DeepMind.

The technical challenges were immense. Google DeepMind's Imagen, Veo, and Gemini models worked in tandem to upscale, extend, and enhance footage without compromising the original performances. Warner Bros insisted that all training data remain proprietary, never flowing into Google's public AI systems.

Even incomplete shots received attention. The Tin Man's frantic axe swings were "outpainted" to expand their cinematic impact, whilst maintaining the scene's emotional authenticity.

Sound Stage to Sphere

Audio presented equally ambitious challenges. The original score was re-recorded on the very sound stage used in 1939, but this time orchestrated for the Sphere's 167,000 speakers. Judy Garland's iconic vocals remain untouched, preserving the film's emotional core whilst surrounding audiences with unprecedented immersion.

Physical effects amplify the experience further. Gigantic helium-filled monkeys sweep across the Sphere's dome, manoeuvred by drone operators as digital counterparts fill the onscreen skies. The boundary between physical and digital dissolves entirely.

Key enhancements include:

• Ultra-high-definition restoration revealing previously invisible details like Garland's freckles
• Outpainting technology expanding incomplete shots for full 360-degree viewing
• Generative AI^✦ models trained specifically on actor likenesses in controlled environments
• Immersive audio design utilising 167,000 speakers for three-dimensional sound
• Physical effects integration including drone-operated flying monkeys

The choice of The Wizard of Oz wasn't accidental. Its original Technicolor transition from sepia Kansas to the vivid Emerald City represented a groundbreaking moment in film technology. Sphere executives recognised symbolic continuity in reimagining a movie that had once set technical benchmarks, similar to how Google's AI tools continue pushing creative boundaries.

Critical Reception and Future Implications

Not everyone embraces this reimagining. Some cinephiles bristle at altering a film that's become part of cinema's canon. Entertainment writer Joshua Rivera called the production "an affront to art and nature." The debate mirrors broader discussions about AI's role in creative industries, as explored in Google's recent AI developments.

Jane Rosenthal, a long-time collaborator of Sphere Entertainment's James Dolan, pushed back against critics. Her perspective centres on audience experience: every change serves to help viewers feel what Dorothy felt. The creative team sees enhancement, not replacement.

Buzz Hays, head of Google Cloud's entertainment industry group, described the project as showcasing how AI can expand rather than erase creativity. Hollywood typically wants to be "the second" to adopt new technologies, he noted. This time, the Sphere forced it to go first.

What makes this different from regular film restoration?

Traditional restoration preserves existing footage. This project uses AI to generate new visual content, expanding scenes beyond their original boundaries whilst maintaining authentic performances and artistic integrity.

How does the Sphere's technology enhance the viewing experience?

The 16K LED dome creates 360-degree immersion with 167,000 speakers and physical effects like wind machines, making audiences feel inside the story rather than watching it externally.

What AI models were used in the production?

Google DeepMind's Imagen, Veo, and Gemini models handled super resolution, outpainting, and performance generation. All training data remained proprietary to Warner Bros, separate from public AI systems.

Will this model be applied to other classic films?

The success could establish AI-enhanced reimaginings as a new norm for studios seeking to reinvigorate classic intellectual property, though each project would require careful consideration of artistic integrity.

How long did the project take to complete?

The collaboration required two years and involved more than 2,000 people, from Google DeepMind researchers to visual effects artists, academics, and studio lawyers establishing responsible AI^✦ use guidelines.