OpenAI Enhances GPT-4 with Improved Image Generation and Safety

OpenAI has introduced GPT-4o, an upgraded version of its AI model featuring native image generation capabilities that allow for the modification of uploaded images and the creation of new ones directly from text prompts. This model demonstrates improved multi-turn consistency and employs an autoregressive generation method for better text rendering and prompt adherence. Additionally, OpenAI has implemented safety measures by including C2PA tags to indicate AI-generated images and restrict content that violates policies, although it will respect explicit user requests for edgier content. Meanwhile, Microsoft has integrated GPT-4 Turbo into its Copilot assistant, granting business users unlimited chat capabilities and improving image generation through Microsoft Designer, utilizing DALL-E 3 for faster creation. In the medical field, GPT-4 has shown promise in assisting neurologists with locating brain lesions post-stroke, demonstrating high sensitivity and specificity, although its accuracy is contingent on the quality of input data. Looking ahead, ChatGPT 5 is anticipated to launch in 2025, focusing on enhanced task automation, reasoning, and language understanding to further narrow the gap between artificial and human intelligence.