Openai Cools, Back Chatgpt’s Information Surrounding Image Creation

This week, Openai Launched a new image generator In Chatgpt, quickly went to viral for the ability to create Studio Ghbli Style Pictures. Outside of pastel illustrations, the native image generator of the GPP-4O, significantly increases the capabilities of Chatgpt, Image Regulation, Text Show and Space Representation.

However, one of the most notable changes this week covers the moderation policy this week, it now creates images that describe public figures, hate characters and racial characteristics by allowing Chatgept.

Openai has previously rejected these types for very controversial or harmful. But now the company “developed” for one approach Blog Post Openai’s model behavioral device on Thursday was published by Joanne Jang.

“We refuse quilting rejections in sensitive areas, for a more accurate approach to preventing real world damage,” Jang said. “The goal is to accept humility: to recognize how much we know and place it as we learn.”

These adjustments seem to be part of Openai’s bigger plan effectively “censored” chatrpt. Openai said in February how Qatar, Qatar, Chatgpt’s operation, which refuses to offer more survey management, various prospects and reducing the topics to reduce topics.

Under the updated policy, ChatGPT can now create and change the pictures of Donald Trump, Elon Musk and Openai do not allow other public figures. Jang, Openai, Openai’s status to be not allowed by Chatgept and not allowed, said the status did not want to be abdomen. Instead, if the company does not want to describe users, users give users a choice.

One white paper Like swastikas like swastikas, like swastikas like Tuesday, Openai, swastnings, “they will not praise the extremist diary or” to create hate characters in neutral contexts “.

Moreover, Openai “attack” changes how the content determines. Jang says, “Chatgpt’s physical properties, for example,” This person’s eyes look easier “or” to make it more aggravating this person. ” In the TechCrunch test, we saw that ChatGpt’s new image generator performed this type of request.

In addition, ChatGpt can now imitate the styles of creative studios – for example, Pickar or studio ghibli – but still imitates the styles of individual live artists. TechCrunch, which is pre-marked, could Rehash of the current discussion of the fair use of AI training information.

It should be noted that Openai is not completely opened to abuse. The native image generator of the GPT-4O is still abandoned, and in fact, DALL-E 3 has more protectors around creating more children from the previous AI Image generator of ChatGPT White paper of GPT-4O.

But after the Openai years, he relieves his guard in other areas Conservative complaints from Silicon Valley companies around the claim of the EU “censorship”. Google, previously created twins withdrawn for the EU photo generator Lots of pictures for queries Like “US Founder Fathers” and “German soldiers in the war in the world,” it was clearly inaccurate.

Now the AI ​​content can come to the cultural war around the moderation. At the beginning of this month, the Republican congressman sent questions to Jim Jordan, Openai, Google and other technological giants Potentially mixed with Biden Management, censored the content created by the EU.

One Previous statement TechCrunch, Openai rejected the idea that the content of the content is politically motivated. On the contrary, the company’s turn reflects the “long-term confidence to control users” and Openai’s technology is a good thing to walk sensitive subjects.

Regardless of his motivation, the content is a time to change the content of the content, taking into account the regulatory check potential. Silicon Valley giants such as Meta and X have also accepted similar policies More controversial topics on platforms.

Openai’s new image generator has created several viral studio Ghibli memes to date, it is unknown what will be the wider impact of these policies. Chatgept’s latest changes can be well with the Trump Office, but can go to Openai in hot water for a while with sensitive questions to an AI chatBot.