Openai Patgpt’s found protective measures around the creation of images

This week, Openai A new image generator is launched in Chatgpt, which quickly passed viral because of its ability to create Ghibli -style imagesS Beyond the pastel illustrations, the native GPT-4O image generator significantly upgrades Chatgpt’s capabilities, improves editing of paintings, depicting text and spatial presentation.

However, one of the most notable changes that Openai made this week includes its content moderation policies that now allow Chatgpt to generate images depicting public figures, hateful symbols and racial characteristics.

Early Openai rejected these types of prompts for being too controversial or harmful. But now the company has “evolved” its approach according to a Blog post Posted on Thursday by the model behavior of Openai, Joanne Jang.

“We are moving from the refusal of the blanket to sensitive areas to a more precise approach focused on preventing damage to the real world,” Jang said. “The goal is to perceive humility: to admit how much we do not know and to position ourselves to adapt as we learn.”

These adjustments seem to be part of the greater Openai plan For an effective “obscene” chatS Openai announced in February that it began to change the way it trains AI models, the best purpose of letting Chatgpt deal with more requests, to offer different perspectives and to reduce the topics that the chat refuses to work with.

According to the updated policy, Chatgpt can now generate and change images of Donald Trump, Elon Musk and other public figures that Openai did not use to resolve. Gian says Openai does not want to be an arbitrator of status, choosing who should and should not be allowed to be generated by Chatgpt. Instead, the company allows users to give up if they do not want Chatgpt to depict them.

In a White paper Issued on Tuesday, Openai also said it would allow Chatgpt users to “generate hateful symbols”, such as Swastikas, in educational or neutral contexts, as long as they do not “clearly praise or approve of extremist programs.”

Openai also changes the way it determines the “offensive” content. Gian says Chatgpt has refused requests about physical characteristics by “make this person’s eyes look more Asian” or “Make this person more severe.” In the Testing TechCrunch, we found that the new Chatgpt image generator is performing these types of requests.

In addition, Chatgpt can now mimic the styles of Creative Studios – such as Pixar or Studio Ghibli – but still limits the imitation of the styles of individual living artists. As TechCrunch noted earlier, this can Restore an existing debate over honest use of copyright in AI training of data setsS

It is worth noting that Openai does not fully open the floods for abuse. The native GPT-4o Image generator still denies very sensitive requests and in fact there are more protective measures about the generation of children’s images than the Dall-E 3, the previous AI generator of Chatgpt AI, according to Chatgpt AI generator, according to Generator The white paper on the GPT-4oS

But Openai releases its railing in other areas after years Conservative complaints about the alleged AI “censorship” by Silicon Valley companies. Google has previously encountered a reverse reaction to the Gemini AI image generator that created Multipricion Request images As “US Fathers” and “German Soldiers in World War II,” which were obviously inaccurate.

Now the cultural war surrounding the moderation of AI content can reach the head. Earlier this month, the Republican Congressman Jim Jordan sent questions to Openai, Google and other technology giants for A potential agreement with the Biden administration for censorship generated by AI content.

In a Previous statement For TechCrunch Openai, he has rejected the idea that changes in content moderate are politically motivated. Rather, the company says that the change reflects “a long-held belief in providing more consumer control,” and Openai technology is only now good enough to navigate sensitive subjects.

Regardless of its motivation, it is certainly an appropriate moment to change its content moderation policies, given the potential for regulatory control under the Trump administration. Silicon Valley giants such as Meta and X also adopted similar policies, allowing More outpatient topics on their platforms.

While the new Openai image generator has just created some memoirs of the Ghibli viral studio so far, it is not clear what the broader effects will be than these policies. The latest Changpt changes can handle the Trump administration well, but leaving AI chatbot answering sensitive questions can land Openai in hot water soon.