As OpenAI recently revealed, DALL·E 3—an upgrade to ChatGPT’s image-generation component—has broadened its capabilities to include the ability to create and alter images from text input. This recent development comes on the heels of widespread alarm over the original version’s flimsy guardrails. Users would easily be able to remove watermarks and create images of copyrighted characters, creating truly egregious concerns. The new feature, leveraging OpenAI’s GPT-4o model, is set to roll out soon to Plus and free users of ChatGPT, as well as developers using OpenAI’s API service.
This expanded and improved image-generation tool comes on the heels of Google testing native image-generation output for its experimental Gemini 2.0 Flash. OpenAI’s latest GPT-4o model “thinking” longer than its predecessor, the DALL-E 3, but guarantees more granular and much more accurate images. Trained on publicly available data and proprietary data from partnerships with companies like Shutterstock, GPT-4o’s native image generation is live today in ChatGPT and Sora, OpenAI’s AI video-generation product.
Subscribers to OpenAI’s $200-a-month Pro plan already have access to this more powerful feature. On Tuesday, OpenAI CEO Sam Altman announced the first big advance in ChatGPT’s image-generating prowess. This is the biggest improvement we’ve made in more than a year!
Brad Lightcap, OpenAI’s chief operating officer, verified the rollout of that new feature in a information to the Journal.
“We’re respecting of the artists’ rights in terms of how we do the output, and we have policies in place that prevent us from generating images that directly mimic any living artists’ work,” – Brad Lightcap, OpenAI’s chief operating officer
The recent addition of GPT-4o including image generation really takes ChatGPT to the next level. The model brings new capabilities that empower creators to design and modify images with more accuracy and intricacy. OpenAI has said that it intentionally tried to avoid these issues while still providing a useful tool that would respect artists’ rights.
What The Author Thinks
OpenAI’s improvement of DALL·E 3 with the GPT-4o model is a significant leap in creative AI technology. By offering more detailed and accurate image generation while addressing ethical concerns about artists’ rights, OpenAI is attempting to strike a balance between innovation and respect for intellectual property. However, the true challenge will be ensuring that these enhancements remain within the bounds of fairness and transparency as they evolve.
Featured image credit: Pond5
Follow us for more breaking news on DMR