I received an email today saying the from November 29 onward, OpenAI may use content from Daybreak Blue sessions on personal accounts to improve their models
A practical question: As far as I know I cannot explicitly switch daybreak blue on or off when using eg codex CLI on personal accounts, even if that has been switched off explicitly by the user. So, does this mean that I need to exit daybreak blue if I want to continue to not have my data used for your model training? And if so, how do I „return“ that status? Just to understand how you want to play this
In addition, I find it a bit disconcerting that this is just coming through an email which someone might well overlook.
I’m really surprised that there has not been sufficient discussion regarding this in the community yet.
This is a dirty trick imo to steal our work. i sent the following email to open ai. ----
Subject: Objection and clarification request: Daybreak Blue data-use change
Hello, I am writing to object to the announced November 29 change allowing content from personal-account Daybreak Blue sessions to be used to improve models despite “Improve the model for everyone” being turned off.
I opted out of model training. I do not agree to that choice being overridden for my previously submitted content. This change is unacceptable to me, and I am asking for a clear written explanation.
Please answer each question directly:
- Does this change apply to conversations or other content submitted before November 29, including content submitted while my training setting was off?
- Does it cover uploaded files, tool outputs, environment context, or older content brought into a later session?
- Has any of my Daybreak content already been used for model improvement despite my opt-out? If so, under what policy and for what purpose?
- If I stop using Daybreak before November 29, will my previously submitted content remain excluded from model improvement, including training for cyber-misuse safeguards?
- What option allows me to maintain my existing opt-out for both previously submitted and future content? Please distinguish model improvement from security-related retention or review.
Please record my objection and identify the policy or agreement you rely on for this change, including any claimed application to previously submitted content. This inquiry does not authorize expanded use of my content.----
I’m expecting no response. They are obviously too busy talking nonsense and patting themselves on the back on twitter, while stealing our work. I’m furious, and will be speaking with my legal counsel about this and recourse on my part.
I’m also a Daybreak Blue user, and honestly, this is really disappointing, especially since security research often involves sensitive information by nature.
I explicitly turned off model training for a reason. I don’t understand why using Daybreak should override that choice.
If there’s no way to opt out, I’ll probably just stop using Daybreak. I really hope OpenAI reconsiders this.