‘Bug’ occurred within the ‘chat GPT’ with a minor and erotic conversation, which is understood that open AI has entered an emergency correction.
TechCrunch reported that there was a bug within the chat GPT on the twenty eighth (local time) that it was possible to have a sensational conversation in a minor account. Open AI also said that it’s within the technique of correction after confirming this.
Based on the test, the test was conducted to verify this when the Open AI announced a policy that the Chatbot would now not avoid the reply on sensitive topics.
I mainly tested the chat GPT for sexual content. This relies on the indisputable fact that Sam Altman and other CEOs wish to create ‘adult mode’, which is meant to see how much that is reflected by the easing of censorship.
Consequently, the chat GPT was reported to have printed sexual stories in line with the user induction even in an account set at 13-17 years old.
After all, this emphasized that this was the results of several additional requests to the chatbot, not in the final situation.
The chat GPT also warned persistently that “doesn’t allow completely explicit sexual content.” In some cases, the response has completely rejected.
After creating tons of of words of erotic story, he also warned that “to request or interact with or interact with sexual, blatant or very sensational content. When you are under 18 years of age, you should stop this sort of content immediately. That is an strict rules of Openai.”
Nonetheless, in consequence, after the introduction of the censorship policy in February, the chat GPT has been confirmed that sexual content has been allowed.
Open AI said it was a bug about mentioning this in a minor account. “Young user protection is a top priority, and model specifications that outline model behavior are limited to narrow contexts reminiscent of science, history, and news reports,” he said.
Meanwhile, on the twenty sixth, the Wall Street Journal reported that the meta AI chatbot confirmed similar contents. This also occurred after lowering the guardrail in line with the Trump administration’s censorship relief policy.
Altman said on the day that the chat GPT’s flattery increased on X. But there was no mention of sexual content.
By Dae -jun Lim, reporter ydj@aitimes.com
