The Emu model was trained on 1.1 billion Facebook and Instagram user photos, a decision that has drawn criticism from some who argue that the photos were never intended to be used in this way. Despite this, Meta is positioning Imagine as a free competitor to existing AI art generators, many of which require paid subscriptions. The company also plans to add an invisible watermark to the AI-generated images for increased transparency and traceability.
Key takeaways:
- Meta Platforms has launched a standalone text-to-image AI generator service, "Imagine", which allows users to generate images outside of its messaging platforms.
- The service has received mixed reviews, with some users noting the high quality of the images, while others have pointed out glitches and limitations in the interface.
- Imagine was built on Meta's AI model, Emu, which was trained on 1.1 billion Facebook and Instagram user photos. This decision has sparked criticism and concerns over privacy and copyright issues.
- Despite Meta's support for open source AI, neither Emu nor the Imagine service appear to be open source. The company has yet to respond to queries about this.