OpenAI Unveils Advanced Image Generation in GPT-4o, Enhancing Creativity and Accuracy

OpenAI recently introduced a new image generating function to ChatGPT, combining its most powerful visual model with GPT-4o. OpenAI CEO Sam Altman defined it as a "incredible technology" that promotes creative power through allowing users to create extremely detailed and contextually precise visuals. The tool aims to achieve a balance between creative freedom and maintaining appropriate limits on offensive content. OpenAI is dedicated to monitoring usage and changing restrictions based on society feedback, especially when it comes to Artificial General Intelligence (AGI).  

In a blog post, OpenAI highlighted the model's main benefits, such as improved text rendering, exact prompt consistency, and higher stability across iterations. Unlike previous AI models, which struggled with precise details and entity associations, GPT-4o can generate photos containing up to 10-20 entities while remaining reliable. Users can modify photographs using normal conversations, making changes to components such as character designs effortlessly.  

To ensure responsible AI usage, OpenAI has included metadata via C2PA in all generated photos, allowing for validation of AI-created visuals. Extra precautions prohibit the generation of harmful or policy-breaking content, such as deepfakes. However, the model still has issues with non-Latin language, inaccurate image cropping, and small variations in complicated designs.  

The tool is now available to ChatGPT customers in the Plus, Pro, Team, and Free tiers, with Enterprise and Education users coming next shortly. Developers will be able to incorporate the tool via API in the coming weeks.

This article is based on information from India Today