Guide to Creating AI Images with ChatGPT

ChatGPT Image Generation Gains Widespread Adoption

OpenAI's AI assistant ChatGPT has possessed image generation capabilities since March 2025. This feature, temporarily restricted for free users, is now accessible without payment for any OpenAI subscription. While generating images within ChatGPT is simple, there are nuances users can benefit from.

Generating Images with Text Prompts

Users can start the process by typing a description of the image they want to create into ChatGPT's prompt bar. Any description containing the phrase 'create an image' is sufficient for the AI to complete the task. Depending on the complexity of the request and whether the user has a paid plan, generating the image can take a few minutes. Last year, OpenAI updated the model that powers image generation to be faster and better at following instructions. Simultaneously, a dedicated 'Images' section was added to ChatGPT's side menu. This section displays all generated images as well as example prompts and styles to try. This parallels the infrastructural developments discussed in the news article OpenAI Explains Codex CLI's Working Principle.

Editing Images Using Existing Photos

Beyond generating images from text prompts, ChatGPT can also modify existing photos or images uploaded by users. To use this feature, one must click the '+' icon to the left of the prompt bar and select the 'Add photo and file' option. Then, the image to be edited is selected, and the desired changes are written as a description. However, it should not be forgotten that photos uploaded to OpenAI servers may be used by the company to train future models. Users can prevent this by disabling the 'Improve the model for everyone' option in the 'Data controls' section of their account settings. Such data processing approaches offer a comparison to the on-device processing trends highlighted in the news Qualcomm Invests in SpotDraft's On-Device Contract AI.

Options for Editing Generated Images

If one is not satisfied with an image generated by ChatGPT, there are two options: requesting a completely new image or editing specific parts of the existing image. In the mobile app, a new generation can be made by clicking the three-dot icon below the generated image and selecting 'Retry'. To edit a specific area, click on the image, use the 'Select area' button to mask the part to be changed, and write the request. A similar masking feature is available in the desktop interface.

Tips for Better Results

Experts note that being as specific and detailed as possible when making requests to ChatGPT leads to better results. It should be remembered that AI models are not deterministic; even with the same prompt, they may not give the exact same response every time. Additionally, it should be considered that ChatGPT may hallucinate, meaning it can produce details that do not exist in reality. Therefore, patience and several attempts may be necessary. The creative and sometimes unpredictable nature of AI, when considered alongside the integration efforts of other AI assistants like in the news Anthropic's Claude Gains Direct Access to Slack and Canva with MCP Extension, demonstrates the expanding use cases in the industry.

Access and Pricing

ChatGPT can be accessed via web browser, desktop applications, or iOS and Android mobile apps. The image generation feature is offered free for anyone with an OpenAI account, but there is a daily generation limit, and free image generation time with the GPT-5 model can be longer. OpenAI provides the ability to generate more and faster images through paid subscription plans. For example, ChatGPT Plus users paying $20 per month can access 'expanded and faster image generation'. These developments can be evaluated in connection with the news Microsoft Announces Powerful New Chipset 'Maia 200' for AI Inference, at a time when specialized hardware efforts for AI processing are accelerating.

It is also noted that ChatGPT cannot create an exact copy of an existing photo or a real-world event due to copyright reasons. The system prefers to offer artistic interpretations inspired by, rather than depicting, real individuals in the face of such requests. For those wanting to follow the latest developments in tech events, the news TechCrunch Disrupt Tickets Are Selling Out Fast: Why You Can't Afford to Miss It This Year contains important information.