ChatGPT’s image generator now creates public figures and controversial content

OpenAI has significantly relaxed ChatGPT’s image moderation policies, allowing users to generate images of public figures, hateful symbols, and content reinforcing stereotypes upon request. This shift, announced alongside ChatGPT’s viral new image generator, represents a dramatic reversal from the platform’s historically strict content limitations and places the burden on famous individuals to opt out of being depicted rather than proactively protecting their likeness.

The big picture: OpenAI published comprehensive moderation policy changes on March 25, detailing areas where filters have been relaxed while others have been strengthened.

• Users can now request ChatGPT to override previous moderation blocks for generating images of famous people, certain symbols, and content that may reinforce stereotypes.
• This follows OpenAI’s February decision to expand ChatGPT’s training data and reduce topics the chatbot refuses to address.

Expanded creative capabilities: ChatGPT can now mimic certain artistic styles, allowing users to generate content inspired by animation studios and creative houses.

• While restrictions on imitating individual living artists remain, users can generate images in the style of recognizable studios like Studio Ghibli, as evidenced by viral remakes spreading online.
• The policy shift fundamentally changes how public figures are treated, requiring them to actively opt out rather than being automatically protected.

Remaining safeguards: Despite the relaxed policies, OpenAI has maintained or strengthened certain critical content restrictions.

• Protections around child safety have been enhanced, with systems designed to detect and block inappropriate content involving minors, including analysis of uploaded images.
• Filters still prevent photorealistic depictions of inappropriate content featuring public figures, extreme propaganda, and potential deepfakes.

Why this matters: The policy changes reflect the ongoing tension between creative freedom and responsible AI development, with OpenAI shifting more control to users while attempting to maintain safeguards against the most harmful content.