OpenAI held a massive ChatGPT-centric event before Christmas, making several key announcements for ChatGPT and other genAI tools it currently offers. The reasoning ChatGPT model o1 and the text-to-video Sora service got public releases. OpenAI also demoed the o3 model coming later this year, released Canvas, and even made ChatGPT available via WhatsApp texting.
What OpenAI failed to do is keep a privacy promise made earlier this year. The company announced a Media Manager feature back in May. The tool is meant to allow creators of all sorts of content to opt out of ChatGPT and other genAI models training.
OpenAI said Media Manager will be available by 2025, but that hasn’t happened. Not only that, but the company might not be actively working on the feature, and there’s no telling when it will come out.
Training models like ChatGPT require lots of data so the AI can learn to predict what the user needs. OpenAI uses all sorts of data from the internet to train its AI, including copyright content from creators it didn’t obtain access to. Several creators sued the company after finding their works were used to train the AI.
Tech. Entertainment. Science. Your inbox.
Sign up for the most interesting tech & entertainment news out there.
Email: SIGN UP
By signing up, I agree to the Terms of Use and have reviewed the Privacy Notice.
Media Manager should be a tool that lets creators opt out of training for all sorts of media content, not just written text. OpenAI described Media Manager as a cutting-edge tool to protect user privacy:
This will require cutting-edge machine learning research to build a first-ever tool of its kind to help us identify copyrighted text, images, audio, and video across multiple sources and reflect creator preferences.
While that sounded great on paper, we have all forgotten about it. OpenAI has not mentioned Media Manager since May, and development for the privacy tool might not have advanced. Says TechCrunch, which talked to people familiar with the project.
The blog has learned that Media Manager was rarely viewed as an important project internally. A former OpenAI employee told TechCrunch that Media Manager wasn’t a priority. “To be honest, I don’t remember anyone working on it,” they said.
Also, a person who doesn’t work at OpenAI but coordinates work with the AI firm said they discussed the tool with the company but haven’t received any updates.
TechCrunch also notes that an OpenAI legal team member involved with Media Manager transitioned to a part-time consultant role.
None of this comes directly from OpenAI. Insiders who shared information might not have access to the bigger picture. But even if TechCrunch’s information isn’t accurate, there’s still the fact that OpenAI has been quiet on the Media Manager feature for half a year.
If it were finished, the privacy tool could have easily been a “12 Days” announcement. At the very least, OpenAI could have updated the timeline.
A more cynical view is that OpenAI is biding its time. Recent reports said that training next-gen ChatGPT versions isn’t working as planned. OpenAI saw delays for GPT-5, and access to better-quality training data might be one factor impacting development. Copyrighted content that makes it into AI training data could be better quality than other types of content. I’m just speculating here.
What I’m getting at is that a Media Manager tool might be released eventually, maybe once the copyright lawsuits against OpenAI are concluded and the company knows exactly what it can and can’t do when training new AI models. Who knows if Media Manager might be needed once that happens. Also, ChatGPT’s GPT-5 upgrade might get out of development and closer to release by then.
Finally, even if Media Manager were available, it would still require some work from creators to flag out copyrighted content that should not train the AI. Despite that, copyright content might be found in all sorts of places on the internet that creators would not flag. OpenAI would crawl those places and the content might still make it into training data sets.
Media Manager delays aside; regular ChatGPT users have privacy settings at their disposal to opt out of having the data in their prompts train the AI.
The post It’s 2025 and OpenAI’s AI tool to protect creators is still missing in action appeared first on BGR.
