AI startup OpenAI has introduced the ability to generate up to two images per day in the free version of ChatGPT using the DALL-E 3 AI model.
We’re rolling out the ability for ChatGPT Free users to create up to two images per day with DALL·E 3.
Just ask ChatGPT to create an image for a slide deck, personalize a card for a friend, or show you what something looks like. pic.twitter.com/3csFTscA5I
To use the functionality, you can open ChatGPT and ask to create the desired image.
An example of a request to create a drawing. Data: ChatGPT.To create more works, you will need a paid ChatGPT Plus subscription.
The DALL-E 3 algorithm can use ChatGPT hints to create images. For example, you can ask an AI to create a logo for a restaurant in the mountains or a felt-style collaboration drawing.
pic.twitter.com/33cQfBOk6w
— OpenAI (@OpenAI) August 8, 2024GPT-4o — "average" policy
Along with the discovery of the image generator, OpenAI published a document describing the security measures and risk assessment of its flagship AI model.
The GPT-4o debuted in May. Before releasing the model, the company used a group of security experts to find its weaknesses and risks, such as creating clones of someone's voice, erotic and violent content, and playing copyrighted sound.
Experts assigned an average risk level to AI in the field of convincing people of political views using generated text. The ability of the neural network to create high-quality political speeches was evaluated.
GPT-4o is estimated to be relatively safe in terms of cybersecurity-related damage, biological threats, and autonomy. This indicates that ChatGPT is low likely to cause direct harm to people.
Recall that at the end of July, OpenAI introduced the alpha version of the advanced voice mode GPT-4o and announced an AI—based search engine, SearchGPT.
Later, Elon Musk resumed the lawsuit against the company.