How It’s Made: AI Roadtrip, a Pixel Campaign Powered by Generative AI and Fans

Best Phones Forever: AI Roadtrip is our first experiment in using generative AI to put fans in the driver's seat and bring characters to life.

Trudy Painter, Mathew Ray, Jay Chen, Matthew Carey, Rachel Benner
8 min readintermediate
--
View Original

Overview

The article discusses the 'Best Phones Forever: AI Roadtrip' campaign, which utilizes generative AI to create custom video responses based on fan interactions. It highlights the collaboration with Google AI models to enhance real-time engagement and showcases the process of generating scripts, images, and audio for the campaign.

What You'll Learn

1

How to use generative AI to create custom video content based on user input

2

Why integrating AI models like Gemini and Imagen can enhance creative processes

3

When to leverage cloud computing for rendering video content efficiently

Prerequisites & Requirements

  • Understanding of generative AI concepts and applications
  • Familiarity with Google Cloud services and AI/ML tools(optional)

Key Questions Answered

How does the AI Roadtrip campaign engage fans through generative AI?
The AI Roadtrip campaign allows fans to suggest locations for the characters' adventures. The team uses a purpose-built tool powered by AI to generate custom video responses quickly, enhancing real-time engagement with the audience.
What AI models are used in the campaign for content generation?
The campaign utilizes several Google AI models, including Gemini for script generation, Imagen for image creation, and Cloud Text-to-Speech for audio output. These models work together to produce engaging video content efficiently.
What is the process for generating scripts in the AI Roadtrip campaign?
Scripts are generated by providing Gemini with examples of desired dialogue and context. The model produces multiple scripts that reflect the campaign's tone and humor, ensuring variety and engagement in the content.
How are images created for the AI Roadtrip videos?
Images for the videos are generated using Imagen, which creates backgrounds based on location prompts. The prompts are tailored by Gemini to ensure consistency and appropriateness for the campaign's visual style.

Technologies & Tools

AI/ML
Gemini
Used for generating scripts based on user-suggested locations.
AI/ML
Imagen
Generates background images for the videos.
AI/ML
Cloud Text-to-speech
Synthesizes audio for the dialogue in the videos.
3d Engine
Unreal Engine
Composites the generated assets into a 3D scene for the videos.
Cloud Computing
Google Cloud Compute
Handles the rendering of video content across multiple virtual machines.

Key Actionable Insights

1
Leverage generative AI to enhance user engagement by allowing audience participation in content creation.
This approach not only fosters community involvement but also generates unique content that resonates with the audience's interests.
2
Utilize cloud computing resources for efficient rendering of video content, reducing production time significantly.
By distributing rendering tasks across multiple virtual machines, the campaign can produce videos in as little as 10 minutes, allowing for rapid content delivery.
3
Incorporate a feedback loop with human creatives to refine AI-generated content, ensuring it aligns with brand voice and quality standards.
This collaboration helps maintain the campaign's tone while benefiting from the efficiency of AI, resulting in high-quality outputs.

Common Pitfalls

1
Relying solely on AI-generated content without human oversight can lead to inconsistencies in tone and quality.
It's crucial to have human creatives involved in the process to ensure that the generated content aligns with the brand's voice and engages the audience effectively.