Abstract:
Change.org, the world-renowned online petition platform, is undergoing a large-scale AI transformation. The company recently revealed that it plans to invest approximately US$100 million of its own funds to use artificial intelligence to rebuild its core petition platform, hoping that online petitions will no longer simply collect signatures, but will further help ordinary people contact decision-makers and promote specific actions.
Ben Rattray, founder and current CEO of Change.org, said that the company hopes to use AI to improve the actual effectiveness of citizens' participation in public affairs. In his view, many people believe that their voices cannot have an impact, and Change.org hopes to change this trust in democratic participation starting from the grassroots.
Change.org started out as a venture-backed company, but following a reorganization in 2021, investors including Reid Hoffman and Sam Altman donated their shares to a non-profit structure. Today, the company says about 500,000 petitions appear on the platform each year, with hundreds of millions of people signing.
This means that Change.org itself already has a large-scale citizen participation network, and the goal of AI transformation is not to simply allow users to generate a petition faster, but to try to make each petition have a more practical political and social impact.
The first phase of testing has already begun. Change.org has launched an AI assistant for petition initiators that can help users draft petition content, find organizations or individuals who may support the petition, and analyze what channels should be used to contact decision-makers who can really solve the problem.
Traditional online petitions tend to follow a relatively simple model: users make demands, others sign, and then the accumulated number of signatures is displayed to government departments, businesses or other institutions. But Change.org believes that simply having a large number of signatures does not necessarily lead to actual results. The real hard part is finding the right decision-makers and getting them to take action.
Therefore, the company plans to use AI to further expand its human service capabilities. The platform will collect and analyze a large amount of customized data to help determine which local government departments, councilors or other decision-making bodies are involved in a certain petition.
For local petitions that reach a certain scale, Change.org even plans to directly intervene. The company said that for local petitions that receive at least 100 signatures, its staff can proactively contact local elected officials on behalf of the sponsor, or even communicate directly with relevant officials by phone.
This means that the role of AI on the Change.org platform is not just "helping you write a petition", but gradually extending to "helping you find who you should contact", and then the platform staff will actually conduct the communication.
From this perspective, Change.org attempts to change the core business and social model of traditional online petitions. In the past, the platform was mainly responsible for providing a space for the public to express their opinions and gather signatures. In the future, it hopes to become an intermediary layer connecting ordinary citizens with specific decision-makers.
However, Change.org does not view AI as a universal tool to solve all problems. Rattray said the company's own tests have found some very interesting results: petitions that rely entirely on AI-generated images and text are less effective than content produced by humans.
A similar situation occurs in videos produced by supporters. The platform found that videos completely generated by AI did not show stronger communication effects.
This shows that in scenarios involving public issues and social mobilization, AI-generated content is not necessarily more convincing than real human expressions. The value of petitions comes not only from the words themselves, but also from the personal experiences, emotions, and social relationships that the initiators and supporters truly invest in them.
Therefore, Change.org is not currently trying to completely replace humans with AI, but hopes to use AI more for information sorting, target identification, and action coordination.
The platform also adopts a relatively cautious attitude towards AI agents participating in petitions. Change.org prohibits users from using AI agents to automatically sign large numbers of petitions for themselves because the company believes that this approach may undermine the most important basis of the petition mechanism - that the participant is indeed a real person and actually supports the relevant demands.
If AI agents can automatically create a large number of identities and sign petitions in batches, millions or even tens of millions of "signatures" may appear quickly on the platform, but these numbers cannot actually reflect real public opinion. This mechanism can also quickly reduce the credibility of online petitions for governments and other decision-makers.
Therefore, Change.org believes that AI can help people express their true demands more effectively, but AI should not be used to create false public participation.
The company also stated that it has established an internal review mechanism to ensure that AI amplification of petition content will not have obvious political leanings. Since the issues dealt with by Change.org often involve political and social issues, the platform hopes to maintain a certain degree of political balance in the AI recommendation and dissemination mechanism.
This is also one of the most noteworthy aspects of this US$100 million investment. Change.org is not simply adding a chatbot to the existing website, but plans to redesign the petition platform from the bottom up, allowing AI to participate in content creation, audience finding, decision-maker identification, and follow-up actions.
If this model succeeds, the meaning of online petitions may change significantly. After users submit petitions in the future, they will no longer just wait for the number of signatures to increase, but will be able to obtain a more complete set of action plans: AI will help sort out the demands and materials, find the most relevant decision-makers, and then Change.org staff will conduct actual communication on behalf of the users.
However, it still takes time to verify whether AI can truly improve the political influence of petitions. Especially when it comes to issues involving complex policies, conflicts of interest, and public resource allocation, AI can help process information, but it cannot replace real political consultations between government agencies, elected officials, and social organizations.
Change.org’s investment of US$100 million is actually conducting a large-scale experiment: if AI is used to expand real citizen participation instead of creating fake traffic, can Internet petitions be transformed from a tool mainly used to express opinions to a public action platform that more directly connects ordinary people and decision-makers.
If the experiment is successful, AI's role here will not only be a content generator, but may become a new "citizen action infrastructure." But Change.org itself has found that human-original content is still more effective than pure AI-generated content, which means that even if AI becomes the core technology of the platform, real human experiences, demands and participation themselves are still the most important components of online citizen actions.
Comments