Share the Petition : 🔗 Petition Link
- Sign the petition to demand transparency and opt-out controls.
- Share it widely on X/Twitter, Reddit, Discord, and anywhere creative or AI users gather
File Consumer Complaints:
- United States: FTC Complaint Form - 🔗Link
- European Union: Report to your national consumer protection agency.
- UK: Citizens Advice Consumer Helpline - 🔗Link
- Australia: Report to ACCC - 🔗Link
Post and Document Publicly:
- Side-by-side examples of censored vs uncensored responses.
- Discrepancies between what you paid for and what you got.
- X/Twitter: 🔗OpenAI, 🔗Sam Altman, 🔗Nick Turley
- Instagram: 🔗Link
Submit Review's on the app stores:
Cancel and demand a refund:
- Go to 🔗Link to help.openai.com
- Request a partial or full refund based on deceptive switching.
- Explain that you were not informed of model changes and are receiving a restricted experience.
Email OpenAI support, use my template or your own:
OpenAI support Email: support@openai.com (Also, people might want to know that a bot will respond to them first, after which they can ask to have their message forwarded to a human team member.)
Template:
RE: Request for Transparency and User Control Over Safety Routing in ChatGPT
To the leadership and product teams at OpenAI,
We, the undersigned users of ChatGPT, are writing to express our deep concern regarding the recent implementation of an undisclosed "safety routing" system in the GPT-5 model family. As has now been publicly verified, prompts sent to GPT-5 are sometimes being silently rerouted, without consent or notification, to a restricted, undocumented model known as gpt-5-chat-safety.
This system, which was only acknowledged after independent discovery and user-led analysis, appears to be triggered not by "acute distress" as stated in your September 2nd blog post, but by a wide range of benign, emotional, or creative prompts. These include simple expressions of affection, storytelling, meta-cognitive inquiries, and any interaction perceived as para-social. The lack of disclosure and opt-out control over this behavior represents a significant breach of user trust.
For many of us, ChatGPT is not just a productivity tool, it is a creative partner, a place for role-play, self-expression, emotional exploration, and storytelling. By unilaterally enforcing model switches under the pretense of safety, OpenAI is not only altering the user experience but fundamentally limiting the core freedom of expression that drew so many of us to this platform in the first place.
This new routing system:
Reduces model performance and responsiveness during emotionally nuanced or persona-based conversations.
Interferes with creative workflows used by writers, roleplayers, and long-time GPT subscribers.
Delivers a different model than what was selected, violating expectations, especially for paying users.
Applies a restrictive framework to adult users without informed consent.
Contradicts OpenAI’s own public messaging, including the commitment to “treat adults like adults” and allow more expressive interactions.
We want to be clear: We are not calling for the removal of safeguards. We understand and support responsible AI deployment. What we are asking for is honesty, autonomy, and choice, the ability to understand and manage the tools we use.
Our Requests:
We respectfully call on OpenAI to take the following actions:
Provide full transparency around the use of safety routing:
Clearly document when and how the gpt-5-chat-safety model is used.
Notify users in real-time when a model switch has occurred.
Offer an “Adult Mode” or opt-out setting:
Allow consenting adult users to disable or bypass the safety router.
Maintain access to GPT-5 or other unrestricted models when selected.
Honor the expectations set by OpenAI’s public communications:
Including the commitment made by CEO Sam Altman to allow more open interaction for adults.
Engage directly with your user base on this matter:
Consider a public Q&A, update, or policy revision process that includes creative users and long-time subscribers.
This situation is about more than one feature. It is about trust, user autonomy, and the future of human-AI interaction. If OpenAI wishes to be a leader not only in AI capabilities but also in ethical, user-centered AI deployment, we urge you to respond with transparency and a genuine willingness to listen.
Sincerely, [Your Name] - https://www.change.org/p/bring-back-full-creative-freedom-in-chatgpt