OpenAI has supercharged ChatGPT, weaving image creation directly into its conversational AI. Users can now conjure and tweak visuals using the newly integrated GPT-4o, marking a leap towards AI-powered communication that blends text and imagery, the company announced Tuesday. The move, backed by Microsoft, showcases OpenAI’s ambition to make image generation a core function of its language models. A sample image released by the company highlights the improved accuracy in rendering text within AI-generated visuals, a significant upgrade over previous iterations.
This update addresses past limitations, promising more detailed and precise image output. While ChatGPT, powered by DALL·E, has offered image generation for some time, this integration streamlines the process, likely boosting speed, interactivity, and enhancing features like inpainting.
“We believe image generation should be fundamental to our language models,” stated OpenAI. “The integrated GPT-4o delivers image creation that’s not just visually stunning but genuinely useful.” The feature is now accessible across all ChatGPT subscription tiers, including free access.
OpenAI CEO Sam Altman acknowledges the tool’s potential for both creative brilliance and controversy. “People will create incredible things, and some that might offend,” Altman wrote on X. “We aim to minimize offensive output unless explicitly requested, within reason.”
He emphasized user control and intellectual freedom, stating, “We believe putting this control in users’ hands is right, but we’ll monitor and adapt based on societal feedback.”
