Creative Commons, the nonprofit organization historically known for pioneering standardized licenses that allow creators to share work while maintaining certain intellectual property rights, has announced a new initiative tailored for the artificial intelligence community. Dubbed “CC signals,” the project provides dataset holders a standardized method to clearly specify how their content can be used in training AI systems.
This new framework is designed to balance the internet’s longstanding ethos of openness with growing pressures surrounding data availability for AI development. Creative Commons said it aims to prevent a situation where entities increasingly restrict access—through paywalls or other barriers—in response to intensive data harvesting by AI companies.
As outlined by the nonprofit, ongoing large-scale data extraction risks eroding open access to content, prompting platforms to impose stricter access policies. Already, notable examples illustrate this tension: X (formerly Twitter) initially permitted third-party AI firms to use publicly available posts as training data, but recently reversed course; Reddit has begun leveraging its robots.txt files to explicitly prohibit automated data scraping intended for AI training; and Cloudflare is exploring methods to monetize automated access or even misdirect bots through intentionally confusing traffic.
In contrast, the CC signals project proposes a structured approach that blends legal guidelines with ethical considerations. Similar in concept to the widely adopted Creative Commons licenses—which have standardized sharing conditions for billions of creative works online—CC signals will offer clearly defined, enforceable standards laying out conditions under which individual data points or datasets can be responsibly included in AI workflows.
“CC signals are designed to sustain the commons in the age of AI,” explained Creative Commons CEO Anna Tumadóttir. “Just as the CC licenses helped build the open web, we envision CC signals contributing to the development of an AI ecosystem grounded in reciprocity.”
Currently in the early stages of development, CC signals’ preliminary design documentation has been shared publicly online and on GitHub for community feedback. Creative Commons plans to launch an early test phase this November, coupled with a series of open town hall meetings aimed at refining the system through public consultation.
