A policy proposal on our approach to deepfake tools and responsible AI
We’re asking for feedback on a proposed Acceptable Use Policy update to address the use of synthetic and manipulated media tools for non-consensual intimate imagery and disinformation while protecting valuable research.
We are proposing an addition to our Acceptable Use Policies on Misinformation and Disinformation to address the development of synthetic and manipulated media tools for the creation of non-consensual intimate imagery (NCII) and disinformation. We have opened a pull request for public comment on the proposed change, which reads as follows:
Synthetic and Manipulated Media Tools
GitHub does not allow any projects that are designed for, encourage, promote, support, or suggest in any way the use of synthetic or manipulated media for the creation of non-consensual intimate imagery or any content that would constitute misinformation or disinformation under this policy.
Why are we proposing this change?
Artificial intelligence (AI) has rapidly progressed beyond the realm of science fiction. One of its most impressive capabilities is the ability to generate realistic looking images, audio, and video, which has substantial benefits for creativity and innovation, but also introduces new vectors of abuse. As we move into a significant year for global elections, we have observed rising concern for the use of AI-generated disinformation such as deepfake videos, images, and robocalls. Additionally, several high-profile cases of the use of deepfake technology to generate NCII have prompted a wave of state and federal legislative proposals. As a code collaboration platform that could be used to share the tools to make synthetic or manipulated media, we must take responsible measures to address these harms while protecting valuable research.
At GitHub, we have always adapted our policies to reflect new technologies and their impacts. This proposed change addresses concerns about the use of technology to create deepfakes for election disinformation, harassment, and other deceptive misuses, while enabling legitimate uses of this technology.
How are we thinking about this?
Our mission is to accelerate human progress through developer collaboration. One way we further that goal is by making a vast amount of source code available to the public. The more source code available, the more the world can learn from it, secure it, and build off it. So, whenever we make changes to our Acceptable Use Policies, we think hard about how our policies will affect the public availability of code on our platform.
There is always some educational value to sharing code, and removing code from the public removes important information that can be used to address critical challenges, understand key concepts, and enable important innovation. For example, security researchers gain enormous value from examining the source code that generates exploits, which allows them to develop tools that can detect malware and create counter measures for it, but such software also has the potential to be abused. We publicly addressed these questions a few years ago when we updated our policies on malware, where we disallow actively harmful projects but allow dual-use content that is shared for the purpose of security research. We want to take a similar approach here.
GitHub is a home for all developers. If you are developing synthetic and manipulated media tools that contribute to the world’s knowledge, we want that research and creative energy here in this community. It is important for this research to happen in the open where others can learn from it. At the same time, we cannot tolerate tools on our platform designed to create harmful non-consensual intimate imagery or whose purpose is to develop synthetic media intended to misinform and spread disinformation. Through enforcing this policy, we will differentiate between harmful misuse and legitimate dual or general use and consider proportionate responses to possible abuse.
Next steps
You can review our proposed Acceptable Use Policy addition in the site-policy repository, where we continually develop and make updates to our policies collaboratively with the community. We invite all stakeholders to comment for the 30-day period from now until May 20, and look forward to learning from and engaging with the broader community on these important topics.
Tags:
Written by
Related posts
GitHub and JFrog partner to unify code and binaries for DevSecOps
This partnership between GitHub and JFrog enables developers to manage code and binaries more efficiently on two of the most widely used developer platforms in the world.
2024 GitHub Accelerator: Meet the 11 projects shaping open source AI
Announcing the second cohort, delivering value to projects, and driving a new frontier.
Introducing GitHub Copilot Extensions: Unlocking unlimited possibilities with our ecosystem of partners
The world of Copilot is getting bigger, improving the developer experience by keeping developers in the flow longer and allowing them to do more in natural language.