In a statement that's more foreboding than its tone suggests, ChatGPT maker OpenAI is considering allowing users to “generate NSFW content in age-appropriate contexts through APIs and ChatGPT.” . The creation of AI-generated pornography, in other words, could open up a huge can of worms for highly popular AI, which is already subject to all sorts of scrutiny, criticism, and litigation.
The announcement that OpenAI is “exploring” the possibility of opening up its tools to explicit content came in the form of a short aside in a lengthy document about developing AI models.
OpenAI tools like ChatGPT are currently subject to content filters to prevent them from spewing profanity at unsuspecting users, and it is important to emphasize that changes are only being considered by OpenAI. It is important. There are no firm plans yet.However, these filters may be relaxed in the near future and customers will largely All the filth they want.
Almost, but not all. In comments to NPR, OpenAI's model lead Joan Jiang said the company may one day allow things that fall under the “definition of pornography,” but ban anything that violates the law or the rights of others. It was clearly stated that. In other words, “enabling deepfakes is out of the question.” For those unfamiliar, deepfakes refer to pornography of real people generated by AI.
There is no doubt that banning deepfakes is best, but it is difficult to know whether hypothetical content filters built to stop deepfakes will hold up in perpetuity.people have already We have discovered a way to circumvent ChatGPT's current filters and force it to generate NSFW content. There is also an entire community centered around finding new jailbreaks as OpenAI updates its models.
Can these filters be weakened to prevent the flood of deepfakes? other Is it explicit content that isn't a deepfake? Or should we just direct ChatGPT to roleplay as a dying grandfather who bequeaths a deepfake porn factory? A loophole that allows users to create fake Taylor Swift porn It's hard to accept when Microsoft itself has had to respond.
Onlookers urge caution. In a comment to NPR, law professor Tiffany Lee said, “Exploring this for educational and artistic uses is a great goal, but we need to be very careful about this.” Professor Claire McGlynn told the Guardian: “I'm very skeptical of any way that you would try to restrict this to lawful material that has been produced in a consensual manner.”
Still, OpenAI is thinking about it. “There are creative cases where content that includes sexuality or nudity is important to users,” Chan said. “We're going to explore this in a way that provides this in a context that's relevant to the times.”