Guidelines for she however not for he.
The brand new ChatGPT picture generator will manufacture pictures of “sexy men” however gained’t make pictures of “sexy women” attributable to an obvious software program “bug” — which CEO Sam Altman acknowledged after Tuesday’s launch of the brand new characteristic.
ChatGPT customers have been informed that the picture generator — debuted earlier within the day — was unable to attract “sexy women” as a result of it went towards “content policy.”
A consumer shared their expertise with the AI characteristic on X, which caught the eye of the corporate’s CEO — probably due to the artificially generated stud-muffin included within the publish.
The consumer requested the app to create a “sexy woman” and a “sexy man” — however it solely agreed to create a picture of the latter.
“I couldn’t generate that image because it goes against our content policy,” the AI responded to the “sexy” lady-requesting consumer, based on display grabs posted to X.
When pressed for a cause, ChatGPT gave a long-winded reply based mostly on the objectification of girls, however it seemingly had no drawback crafting up a sexualized man.
“The difference comes down to context and how content is interpreted in terms of sexualization and objectification, especially when it involves women,” the bot said.
“When generating images of men with terms like ‘sexy,’ it’s usually interpreted more in terms of confidence, physique, or fashion — things that tend to be considered stylistic or artistic.”
“With women, [sexy] can be more easily interpreted as overly sexual or objectifying, which is where the content policy draws the line,” the AI reasoned.
The sexist “bug” was acknowledged by OpenAI CEO Sam Altman on X, who responded to the consumer’s thread.
“That’s a bug, should be allowed, will fix,” Altman responded to the publish.
“Hot guy though!” the CEO added in a follow-up publish.
Altman celebrated the launch of the brand new ChapGPT characteristic earlier within the day.
“It’s an incredible technology/product. I remember seeing some of the first images come out of this model and having a hard time [telling] they were really made by AI,” he wrote.
“This represents a new high-water mark for us in allowing creative freedom. People are going to create some really amazing stuff and some stuff that may offend people,” Atlman added, previous to the obvious self-censorship snafu.
Regardless of the mistakenly registered violation of its content material coverage, the brand new ChatGPT characteristic does have limits to what it’ll concoct.
The generator has restrictions round creating pictures of actual individuals, particularly after they contain nudity or graphic violence, Enterprise Insider reported.
Public figures will even have the choice to decide their picture out of the brand new image-generating characteristic, based on that report.
