A whole new update has just been deployed on Chatgpt. Artificial intelligence now promises an image generation without restrictions, but this functionality is experiencing some turbulence.
Chatgpt marks a big step in the generation of images. Formerly unconvincing and obsolete, the clichés generated by the artificial intelligence of Openai are experiencing this Tuesday, March 25, 2025 an important update. Competition with Midjourney and Aurora is relaunched.
The new chatgpt update, although it brings a long -awaited improvement (which is more free) of its image generation tool, meets some turbulence. “The images in Chatgpt are much more popular than we had planned (and we had fairly high expectations)”admits Sam Altman, the CEO of Openai, on X. “The deployment of our free version will unfortunately be delayed for a while”. No date has been mentioned.
An improved image generation tool
Chatgpt welcomes this March 25, 2025 a whole new extension for GPT-4O, its free multimodal model deployed in May 2024. The image generation tool developed by OpenAI, soberly baptized “4o image generation”, promises a better understanding of requests and significantly improved performance.
“We are launching something new today: images in Chatgpt!”exclaimed Sam Altman, CEO of Openai, on X (formerly Twitter). “It’s incredible technology […] We believe that people will love it, and we are impatient to see the resulting creativity “.
GPT-4O can now generate images loyal to user demand. Openai’s artificial intelligence is in fact capable of understanding between ten and twenty information in a single prompt. Much better performance than Dall-E, the Chatgpt images generation program, which can only integrate three or four.
GPT-4O also has an image editing capacity. Users can submit the cliché of their choice to see it modified according to their request. The chatgpt model has a “memory” and retains without problem up to 15 photos, useful when you want to submit to AI references or inspirations.
A lack of moderation concerning
Promising, the chatgpt update is to be taken with tweezers. For the sake of realism and freedom of expression, Sam Altman announced on his X account that his AI could now generate known faces of the general public: “This represents a new summit for us in terms of creative freedom. People will create really incredible things, but also things that may offend people.”
Sam Altman, however, encourages chatgpt users to remain respectful: “What we would like to target is that the tool does not create offensive content unless you want, in which case it will do it within the limits of reasonable […]. We will monitor how things will take place and listen to the community “.
It nevertheless seems obvious that this new freedom of expression incorporated on Chatgpt will bring several drifts. The lack of moderation makes the generation of offensive and humiliating images possible and multiplies the risk of disinformation. Fortunately, a few guards have been set up with the update, including the ban on generating erotic images or pornographic deepfakes.
New chatgpt features are accessible for everyone for free. If the update has been deployed on March 25, subscription -free users will have to wait a little longer. Victim of its success, the update will surely be applied for several weeks.