Hello,

Now that the latest version of ChatGPT is out, OpenAI has received a lot of publicity touting all the new features. There’s just one important feature still missing: transparency. We still don’t know how OpenAI trains its AI models, and what data it uses to do so.

To be clear, we don't have basic information about how the data is collected or filtered in their AI tools. OpenAI has dodged questions from Mozilla and others – but we are not giving up because all of us deserve to know how this new technology was built. And if OpenAI refuses to provide answers, it makes us more convinced that they may be hiding something.

OpenAI isn’t the only company that needs to be more transparent about its AI tools – Mozilla is also calling on Microsoft and Google to do the same. But right now, OpenAI is a high-profile leader of this new era of technology, and what they do has a large impact on AI advancements as a whole.

Demanding transparency from powerful and irresponsible tech companies about their AI models is not going to be easy – it will require a huge effort in order to keep up the pressure. And right now, it is up to the Mozilla community to use our collective power to ensure the next wave of AI is for the public good.

Mozilla's campaign work is made possible by tens of thousands of grassroots donors who have contributed to reclaim the internet. Today we are hoping you'll join them by adding your donation today.

Contribute €23 to Mozilla today so that together we can demand transparency from powerful and irresponsible tech companies and ensure AI is trustworthy.

OpenAI has seen lots of controversy over the last few months, with its board trying to oust CEO Sam Altman over a “breakdown in trust” and the recent departures of leading staff involved in AI safety efforts. But the controversy doesn’t seem to have changed OpenAI’s attitude to how it develops AI. As CEO Sam Altman said earlier this month, “I think that you can just do stuff in the world… You don’t need to wait, you don’t need to get permission.”

Take the recent example of actress Scarlett Johansson. When OpenAI released ChatGPT with a voice that closely resembled her own, Johansson's lawyers sent inquiries requesting them to detail the exact process by whichOpenAI had created the voice. After not providing any detail on how the voice was developed, OpenAI quickly removed it from the latest ChatGPT version.

How can we confront the problems that generative AI is causing? It starts with knowing how companies train their AI models, and what data they use. We currently have a limited grasp of how different companies' technologies work and why AI models behave in certain ways – which is why it’s past time for better transparency.

The Mozilla community is taking on some of the most powerful tech companies in the world, and it’s important that all of us stay involved. Mozilla is conducting research on AI, running campaigns, and commissioning independent research projects to scrutinize AI models and document issues with AI, such as amplifying racist content. Because the real-world impact of these AI tools is too great for them to remain largely opaque.

So if you agree it's time for OpenAI to be held accountable for how they build AI models, we hope you'll consider making a contribution today.

Contribute €23 to support Mozilla's campaign efforts to take on powerful and irresponsible tech companies, demand transparency on AI models, and make AI more trustworthy.

Thank you for all you do for the internet.

Michael Whitney
Director, Digital Engagement
Mozilla