An icon of an eye to tell to indicate you can view the content by clicking
Signal
Original article date: May 08, 2026

Sony's AI Playbook: How PlayStation Is Using Generative AI to Augment Game Development

May 11, 2026
5 min read

Sony is drawing a clear line on AI: use it to work faster, not to replace the humans who make games great. During an earnings presentation on May 8, 2026, PlayStation outlined its emerging AI strategy — one built around augmentation, not automation of creative work.

AI as a Studio Tool, Not a Shortcut

Sony's position is that AI is a "powerful tool," but the company was explicit that "the vision, the design, and the emotional impact of our games will always come from the talent of our studios and performers." The message is deliberate: AI is there to help developers work better, not to replace them.

Practically, that means PlayStation studios are using AI to:

  • Automate repetitive workflows and improve software engineering productivity
  • Accelerate quality assurance, 3D modeling, and animation production
  • Reduce friction in performance capture processing through tools like "Mockingbird"

Mockingbird: A Real-World Example

One of the most concrete examples Sony shared is an internal tool called Mockingbird, which animates 3D facial models using performance capture data — completing work that previously took hours in a fraction of a second. Naughty Dog (The Last of Us) and Santa Monica Studio (God of War) have both used it, and the tool has contributed to released titles including Horizon Zero Dawn Remastered.

Sony's caveat: "We are not replacing human performers, but rather optimizing how we process the data from these live captures."

The Bandai Namco Partnership

Sony also announced a collaboration with Bandai Namco to "explore how Generative AI and the latest technologies can most effectively contribute to realizing a creator's vision in the realm of video production." Early results show "massive gains in speed and productivity per person," though Sony also flagged the "lack of consistency and controllability" as a known weakness of generative AI models.

🔗 Read the full article on The Verge