Canva unveils AI 2.0 at Los Angeles conference to redefine digital creation

Canva has unveiled Canva AI 2.0 at its flagship "Canva Create" conference, marking the culmination of a two-year internal overhaul designed to reposition the company as a machine-learning-driven platform.

Announced at SoFi Stadium on Thursday, the update signifies a shift away from being a technology-supported design tool toward becoming a unified, agentic creation system.

The core of the transformation lies in its "object-based intelligence," which co-founders Cliff Obrecht and Cameron Adams highlighted as a major advancement over previous flat-output models.

The system now allows the AI and users to modify individual layers and components separately. This means users can adjust a single headline or swap a specific image without disturbing the rest of the design layout.

Furthermore, when users upload their own content, the system automatically analyses and deconstructs it into editable components.

A significant focus has been placed on "Living Memory," a persistent system that learns from the activities of individuals and entire organisations.

Teams now have the authority to manage these memory files, ensuring the AI adapts to their specific working styles while strictly adhering to established brand guidelines.

This combination of personalisation and restriction aims to maintain consistency across large-scale content creation.

Currently, Canva AI 2.0 is available as a research preview. Reflecting the company's playful brand identity, access is restricted to the first million users who discover a hidden "Easter egg" password revealed during the keynote. This rollout represents Canva's most significant evolution since its inception in 2013.