A Comprehensive Guide to Generative AI for Civil Society

The Growing Concern of Generative AI in Civil Society Organizations

The rise of generative AI technologies, including tools like ChatGPT and Gemini, has sparked discussions among civil society organizations (CSOs) about the appropriateness and implications of these tools in their work. An alarmingly common scenario many CSOs face is the sight of their activists drafting official statements or communications using these AI services without any formal organizational policy guiding this practice. The absence of such guidelines can lead to confusion, inconsistency, and potential reputational damage.

The Necessity of Organizational-Level Policy

Activists often rely on their judgment when incorporating AI tools into their workflows. However, without a cohesive organizational policy on generative AI, issues can emerge that extend beyond individual discretion. For instance, factual inaccuracies—commonly referred to as "hallucinations"—can find their way into official documents, jeopardizing the credibility of an organization. Incorporating flawed information can mislead stakeholders and the public, damaging trust built over years of community engagement.

Moreover, reliance on commercial AI services can expose sensitive organizational data. Uploading personal or confidential information to unprotected platforms poses significant security threats. This potential for data breaches highlights the urgent need for policies that not only articulate the acceptable use of these tools but also establish protocols for information management and security.

Addressing Bias and Ethical Considerations

Another critical factor is the bias inherent in many AI systems. Generative AI tools are trained on vast datasets that may harbor cultural biases. If organizations unwittingly adopt an AI-generated message that contains bias, it may conflict with their core values and mission. This concern is magnified in contexts where language, cultural perceptions, and ethical considerations are at play. Organizations must be vigilant in reviewing AI-generated outputs to ensure they align with their principles.

Additionally, AI tools might inadvertently skip over essential aspects of organizational deliberation and capacity building. Relying heavily on AI for drafting statements can stifle the important discussions and internal reflection that characterize activist work. Engaging in the process of crafting communications is vital for team cohesion and strategic clarity, elements that can be undermined if individuals default to AI-generated content.

The Korean Context: An Ongoing Challenge

In Korea, the absence of comprehensive guidelines on the utilization of generative AI by civil society organizations raises significant concerns. As these technologies rapidly evolve, it becomes imperative for activists and organizations to establish clear principles governing the use of AI. This gap is especially troubling given the lack of documented insights regarding which tools activists are using for varying tasks, and the perceived utility of these tools.

Understanding that many organizations struggle with this void, a recent initiative sought to assess the landscape by surveying activists. The survey explored the AI tools currently in use, their perceived effectiveness, and the challenges users face. Feedback was sourced not only from local activists but also from the broader APC network, demonstrating the global relevance of these issues.

Workshops and Community Engagement

To build a more nuanced understanding of generative AI’s implications, workshops were organized with activists from different sectors, including civil society and labor unions. These sessions provided a platform for participants to dissect the survey findings and discuss a preliminary policy framework. Importantly, the focus was not solely on reaching consensus, but rather on fostering an environment of honesty in sharing concerns and experiences. This dialogue underscored the reality that the development of policies around AI must reflect the diverse voices within each organization.

The Divergent Perspectives on AI Use

While some activists view generative AI as a valuable asset, many others remain apprehensive about its implications. Concerns surrounding the centralization of power in big tech companies—that dominate the development and provision of generative AI services—only add to this unease. This guide is not intended to pressure organizations into adopting generative AI; rather, it acknowledges the complex landscape in which these technologies operate and the ethical dilemmas they pose.

Ultimately, the aim of this guide is to serve as a resource for organizations and activists contemplating their stance on generative AI. As the conversation around these tools evolves, it is critical for each organization to reflect its unique reality and the perspectives of its members in developing tailored policies.

In this fast-paced digital age, we recognize that while challenges abound, thoughtful engagement with generative AI can lead to informed decision-making, enhanced collaboration, and ultimately, a stronger civil society.