Human-Content-to-Machine-Data_Final - Flipbook - Page 19
But, the absence of someone having a legal right to veto something does not mean their
expectations and preferences do not matter.
Ethically, violation of creator expectations at scale is a problem. It is also short-sighted. Just
as a homeowner does not have to invite you back if you refuse to clear your plate after dinner,
or an employee can look for a new job if they believe their employer9s policies are unfair,
creators can choose not to share with the public at all. If we want a resilient ecosystem for
knowledge sharing, we need creators to consent to their participation in the AI lifecycle.
It9s clear that AI developers need this, too. Many examples of new, generative AI-powered
tools decline to answer queries to avoid accessing news content they are permitted to
access, while answering queries using content they shouldn9t be accessing.76 As we9ve written
before, responsible technology developers share an interest in deûning better ways to
respect creators9 wishes.77 If they could get a clear signal of the creators9 intent, then they
would follow it.
To be sure, the concept of consent is complicated and may sometimes come into conüict
with other values. There are important scenarios where it is both legally and ethically
acceptable to use something, even where the creator doesn9t consent and might outright
object, such as for parody or critical review. The boundaries of consent are not
static—instead, they need to be constructed and developed over time.
A Thriving Commons Requires Reciprocity
We believe that a critical ingredient to widespread consent is reciprocity.
Sharing knowledge generally—and CC-licensed content speciûcally—has always been based
on an implicit assumption that we are all in this together. The basic value system of the
commons is rooted in a fundamental belief that knowledge and creativity are building blocks
of our culture, rather than just commodities from which to extract market value. We give, we
take, and we give again.
This notion is not new to the commons. Scholar Elinor Ostrom described the reciprocal
exchange of trust and cooperation as the key to successful governance of the commons.78
76
Jazwinska, K., & Chandrasekar, A. (2025, March 6). AI Search Has A Citation Problem. Columbia
Journalism Review.
https://www.cjr.org/tow_center/we-compared-eight-ai-search-engines-theyre-all-bad-at-citing-news.ph
p
77
Stihler, C. (2023, August 31). Exploring Preference Signals for AI Training. Creative Commons.
https://creativecommons.org/2023/08/31/exploring-preference-signals-for-ai-training/
78
Ostrom, E. (1990). Governing the Commons: The Evolution of Institutions for Collective Action.
Cambridge University Press.
19
