OpenAI announced that it will start giving up to two DALL-E 3 images for its free users, extending its service from its initial exclusivity to ChatGPT Plus users.
DALL-E 3 launched last year and was only available to Plus users, allowing them to generate AI photos based on prompts.
OpenAI Extends DALL-E 3 to ChatGPT Free Users
OpenAI has been updating its features over the previous months, however, many of its most advanced tools are hidden behind a paywall. ChatGPT Free is often utilized for basic writing assistance and accessing GPT-4o mini.
In an X post, OpenAI shared that it will start offering up two images each day that can be used by free users. A notice of daily limit would appear to users after exceeding two images.
DALL-E can generate images based on specific textual prompts. Some of its examples include creating visual pegs based on a specific theme or a visual representation of what the user might be looking for.
OpenAI Releases Safety Assessment for GPT-4o Model
OpenAI also shared the GPT-4o System Card which detailed the safety measures and risk assessment that the startup conducted before releasing the model to the public. The AI company hired external security experts who explored potential weaknesses of the system.
The experts found that GPT-4o is a "medium risk" in terms of persuasion, while it scored low in terms of cybersecurity, biological threats, and model autonomy.
"Building on the safety evaluations and mitigations we developed for GPT-4, and GPT-4V, we've focused additional efforts on GPT-4o's audio capabilities which present novel risks, while also evaluating its text and vision capabilities," the company explained.
The company's release of a safety assessment for its latest model came at a crucial time wherein several employees voiced their concerns about the safety standards within the company.
Related Article : More OpenAI Lead Members Leave AI Startup Amid Growing Safety Concerns