ChatGPT Image Generation: Now Free for Everyone!
OpenAI's integrated image generation feature within ChatGPT is now accessible to all users, even those on the free tier. This follows a delay announced by CEO Sam Altman, who cited unexpectedly high popularity as the reason for initially restricting the rollout.
Previously, this capability was exclusive to Plus, Pro, and Team subscribers. Now, free users can generate images directly within ChatGPT, eliminating the need to switch to DALL-E.
Usage Limits and Creative Applications
While Altman initially mentioned a limit of three images per day for free users, initial tests reveal some accounts are restricted to just one. Generation times can also vary, ranging from a few minutes to several hours.
One popular application has been the creation of Ghibli-style images from real-life photos. Users upload an image and instruct ChatGPT to transform it into a Ghibli-esque artwork.
Ethical Considerations and Copyright Concerns
This trend has reignited discussions regarding the ethical implications of using copyrighted material as training data for AI. It also brings back Hayao Miyazaki's reaction to AI-generated animation, where he stated, "I strongly feel that this is an insult to life itself."
Ongoing Development and Potential Issues
Altman has acknowledged that the tool may exhibit unpredictable behavior and may incorrectly refuse some image generation requests. OpenAI is actively working to address these issues. The company also stated that it will "take action" against images that violate its guidelines, although specific details about these actions were not disclosed.
Creating images is very simple: users only need to describe what they want, specifying aspects such as aspect ratio, colors and background.
Source: Engadget