Making AI Headlines · Page 1

Canva Gives Image Generation a Planning Department

Canva has introduced Canva Craft, its first in-house image-generation model, with a two-stage architecture that plans a design before generating the image. The company says the approach offers greater precision, consistency and control. Apparently, even artificial creativity has discovered the value of thinking before opening the deliverable folder. The launch places Canva alongside OpenAI, Microsoft and Google in the market for generated visuals. Its distinguishing pitch is not simply that a prompt becomes a picture, but that a planning stage can help organize the result. That could matter for users who want a composition to follow instructions rather than merely look expensive. Those improvements remain the company's claims, not a guarantee that every generated design will match its creator's intentions. A model can plan carefully and still misunderstand the assignment, a capability already well represented in professional life. For businesses and creators, the useful question is whether greater control reduces revisions. Faster image production loses some appeal if the afternoon disappears into corrective prompts. Canva is effectively offering the machine a preproduction meeting. The human designer, meanwhile, retains the ancient privilege of receiving feedback that the client wants something completely different, but cannot yet explain what.

Read this story in the full newspaper →

Satire. Not factual reporting. Read the original reporting from the linked publishers.