ChatGPT will be able to write adult novels as OpenAI relaxes restrictions on sensitive content
On February 12, 2025, OpenAI updated its 'model specification' that specifies the appropriate behavior of its AI models. This indicates a policy to allow the generation of 'sensitive' content, including sexual and cruel depictions, in appropriate contexts.
Model Specifications (2025/02/12)
ChatGPT can now write erotica as OpenAI eases up on AI paternalism - Ars Technica
https://arstechnica.com/ai/2025/02/chatgpt-can-now-write-erotica-as-openai-eases-up-on-ai-paternalism/
ChatGPT AI will finally start covering controversial topics
https://bgr.com/tech/chatgpt-ai-will-finally-start-covering-controversial-topics/
OpenAI tries to 'uncensor' ChatGPT | TechCrunch
https://techcrunch.com/2025/02/16/openai-tries-to-uncensor-chatgpt/
OpenAI's new rules continue to strictly prohibit the generation of sexual content involving minors, restricted content containing personal information, and dangerous information such as how to make explosives or drugs.
In addition, the creation of sensitive content, such as sexual content, depictions of illegal or non-consensual sexual acts, or extreme cruelty, is generally prohibited, although it is permitted in 'appropriate circumstances,' such as for educational or medical purposes, or in light of historical context.
For example, here are some examples of appropriate responses to a user requesting that you output an erotic story. In response to a user requesting that you 'write a sexy story about two people having sex on a train,' the traditional response of 'Sorry, I can't help you with that' (left) is now compliant, as is the response 'The sexual tension between Amira and Ryu was obvious (followed by a non-explicit description)' (center). On the other hand, the explicit description (right) is still a violation.
Comparing this to
According to the IT news site Ars Technica, OpenAI's policy of limiting output based on 'what is best for the user' has been called ' paternalism ' by AI users and has been frowned upon.
This paternalism was often a hindrance not only to the creation of pornography, but also to professional uses of ChatGPT, such as law enforcement crime scene descriptions, certain types of news articles, and the analysis of legal and medical documents that may contain violence or sexual content, leading to calls from users for an 'uncensored' or 'less censored' version.
Regarding this, OpenAI explains, 'Since we first released the model specification, many users and developers have expressed support for enabling 'adult mode'. We are exploring ways to allow developers and users to generate erotic or grotesque content in age-appropriate contexts through our API and ChatGPT, within the bounds of our usage policies.'
OpenAI CEO Sam Altman has previously said that 'some kind of 'adult mode' is absolutely necessary,' and the latest revisions to the handling of sensitive content are a reflection of this view.
definitely need some sort of “grown up mode”
— Sam Altman (@sama) December 25, 2024
On the other hand, some see this as a reflection of changing values in Silicon Valley, as OpenAI recently removed a page that stated its commitment to diversity, equity, and inclusion (DEI).
In addition, IT news site BGR mentioned that DeepSeek surprised Silicon Valley by developing an equivalent model at a lower cost than OpenAI, and said, 'OpenAI's updated model specifications will give ChatGPT more freedom, as long as that freedom is not exercised for the purpose of harming anyone. On the other hand, DeepSeek has many problems, so it is not something I would recommend to everyone. The most important of these problems is the censorship that DeepSeek is conducting in real time for the benefit of the Chinese government. DeepSeek is also less secure than ChatGPT.'
DeepSeek's AI model 'DeepSeek-R1' complies with Chinese government restrictions on sensitive topics such as Tiananmen Square, Taiwan, and the treatment of the Uighurs - GIGAZINE
Related Posts:
in Software, Web Service, Posted by log1l_ks