Cook County Government

The Global Dialogue on AI Governance presents a critical opportunity to address an often-overlooked dimension of artificial intelligence: the preservation and integrity of historical archives. Every nation holds a complex historical record shaped by both achievement and injustice. These archives are not merely repositories of the past; they are foundational to accountability, collective memory, and informed global cooperation. As AI systems increasingly mediate how information is stored, retrieved, summarized, and disseminated, they introduce new risks to the authenticity, completeness, and accessibility of these records. Of particular concern is the potential for historical erasure, whether through deliberate manipulation, selective omission, or algorithmic bias. Changes in political leadership or ideology must not be allowed to distort the historical record, especially when AI systems can amplify or obscure narratives at scale. Without safeguards, AI may inadvertently enable the rewriting of history in ways that undermine truth and reconciliation. For example, the UN resolution condemning the transatlantic slave trade, along with each member state's recorded position, must remain permanently and transparently accessible, as these records are essential to historical accuracy and ongoing global dialogue on justice, reparations, and human rights. I therefore urge the international community to prioritize the protection of archives within AI governance frameworks. This includes: • Establishing global standards for the preservation, verification, and traceability of historical data used by AI systems • Ensuring that AI models do not overwrite, distort, or selectively omit historically significant information • Supporting independent, internationally coordinated archival institutions resistant to political pressure • Promoting transparency in how AI systems source and represent historical records Safeguarding humanity's collective memory must be recognized as a core pillar of responsible AI governance. Without a shared commitment to preserving truth, the benefits of AI risk being overshadowed by its capacity to obscure the very history that informs our global future.

• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Protection and promotion of human rights
• Transparency, accountability, and human oversight

From my perspective, the most urgent priorities identified in General Assembly Resolution 79/325 are safe, secure and trustworthy AI; social, economic, ethical, cultural, linguistic and technical implications of AI; protection and promotion of human rights; and transparency, accountability, and human oversight. These areas are essential to safeguarding the integrity of historical archives in the age of AI. As reflected in my submission, AI systems increasingly mediate how information is stored, retrieved, and presented, creating new risks to the authenticity and completeness of historical records. Ensuring AI is safe, secure, and trustworthy is foundational to preventing manipulation, distortion, or loss of archival data. The broader societal implications of AI are equally critical. Historical archives reflect diverse cultural, linguistic, and social realities. Without careful design and governance, AI systems may marginalize certain narratives or perpetuate bias, undermining the representation of underrepresented communities and distorting collective memory. The protection and promotion of human rights directly intersect with the preservation of historical truth. Access to accurate records, including documentation of past injustices, is essential for accountability, reconciliation, and justice. The example of the UN resolution condemning the transatlantic slave trade highlights the importance of preserving both outcomes and the recorded positions of member states. Transparency, accountability, and human oversight are necessary to ensure that AI systems handling historical data remain auditable and resistant to political or ideological interference. Clear standards for sourcing, documentation, and traceability are critical to maintaining trust in AI-mediated knowledge systems. Together, these priorities reflect an urgent need to embed archival integrity within AI governance and to ensure that historical truth remains protected for future generations.

One emerging issue not fully captured in the listed themes is the need for global, AI-supported preservation of digital and physical heritage at risk of loss, erasure, or destruction. Current discussions focus largely on data governance, but less attention is given to what should be preserved and how AI can proactively safeguard humanity's shared record. Archives are not limited to documents. They include cultural landmarks, oral histories, endangered languages, and sites of historical significance that may be vulnerable to conflict, climate change, neglect, or political revision. AI presents a unique opportunity to create globally coordinated archival systems that go beyond passive storage. Technologies such as large-scale digitization, geospatial modeling, and multimodal AI can document landmarks, reconstruct damaged sites, and preserve cultural artifacts in ways that are accessible across borders. For example, AI-driven 3D modeling could preserve at-risk heritage sites, ensuring they remain part of the global record even if physically altered or destroyed. At the same time, this raises governance challenges around ownership, representation, and authenticity. Questions of who controls these digital archives, how they are interpreted, and whose perspectives are included require international standards and oversight. The United Nations is uniquely positioned to serve as a coordinating body for this effort, helping to establish shared principles and accountability mechanisms. I would welcome the opportunity to contribute to this work and to help shape inclusive, globally representative approaches to preserving humanity's collective record. This issue connects technology, culture, human rights, and long-term global memory. It also shifts the conversation from reactive protection to proactive preservation. Addressing this gap would mean recognizing preservation as a shared global responsibility and leveraging AI not only to manage information, but to ensure that the full scope of human history, both tangible and intangible, is documented, protected, and accessible for future generations.

In the United States, current governance gaps and recent policy reversals are creating serious risks across all four thematic areas I identified, even as they generate narrow opportunities for innovation leadership. First, the revocation of the prior "safe, secure, and trustworthy AI" executive order removed an emerging federal framework for safety testing, incident reporting, and safeguards around high risk systems. This has widened the gap between rapid deployment of powerful models and meaningful oversight, increasing the likelihood of harmful uses in surveillance, critical infrastructure, and democratic processes without clear accountability. Second, the shift toward "removing barriers" to AI governance has weakened momentum for addressing the social, economic, ethical, cultural, linguistic, and technical implications of AI. In practice, this means less federal attention to bias, discrimination, and the disproportionate impact of AI on marginalized communities, including those whose histories and experiences are already underrepresented in datasets and archives. The current federal administration has recently and intentionally stopped updating key federal records related to Underground Railroad sites and other locations central to African American history and has removed crucial datasets documenting African American communities and other non European immigrants. This is not a neutral omission but an active narrowing of the historical record that AI systems will inherit and legitimize. Third, by deprioritizing multilateral AI governance efforts and framing international regulation as a constraint on "American values," the current approach undermines global cooperation on human rights centered AI norms. The renewed withdrawal from bodies such as UNESCO further weakens international collaboration on cultural preservation, education, and shared standards for safeguarding historical memory. Finally, gaps in transparency, accountability, and human oversight are widening as powerful systems are deployed in sensitive domains without binding federal standards for documentation, auditing, or redress. This absence serves as a major governance risk with global implications.

The AI Dialogue can provide a rare, structured space where states, experts, and civil society confront the political realities of AI, not just its technical risks. It can foreground how AI affects historical memory, archives, and cultural heritage, ensuring that preservation and truth are treated as governance priorities rather than side issues. By convening diverse stakeholders, the Dialogue can help develop shared principles for data provenance, archival integrity, and protections against historical erasure, including in training data and model outputs. It can also serve as a venue to document and challenge harmful practices, such as state-led manipulation or suppression of records, and to elevate the experiences of communities most at risk of being written out of AI-mediated history. Finally, the Dialogue can help bridge fragmented national approaches by encouraging interoperable standards for transparency, accountability, and human rights that travel with AI systems across borders.

Existing initiatives the AI Dialogue should build on include local government innovation efforts, independent policy organizations, and multistakeholder training programs focused on responsible AI. Local governments, especially in countries where national administrations have adopted harmful or regressive AI and data practices, are often closer to affected communities and better positioned to pilot rights-respecting, accountability-focused governance models. As a local government practitioner, I can contribute practical experience from this level of implementation. Independent organizations such as the Center for AI and Digital Policy (CAIDP) and Aspen Digital/ Aspen Policy Academy offer deep expertise on AI policy, human rights, and democratic safeguards. In my role as a Team Lead with CAIDP and a Rising Civic AI Leader with Aspen Digital/Aspen Policy Academy, I see significant potential for the AI Dialogue to connect UN processes with these communities of practice. The added value of the Dialogue would be to globalize lessons from these initiatives, elevate local and civil society perspectives, and embed them into interoperable, internationally recognized AI governance standards.

Civil society cannot be treated as peripheral to this process. In this moment, its expertise and lived experience are too often sidelined in favor of large technology firms and industry voices. A legitimate AI Dialogue must center those closest to harm, not just those closest to power or profit. Different stakeholders can strengthen the AI Dialogue by contributing in ways that match their proximity to impact, their expertise, and their accountability. Governments should provide clear, public explanations of their AI policies, including how they handle data, archives, and human rights, and commit to independent review. Local governments should be explicitly included, especially where national governments pose risks to ethical and responsible AI; they can share grounded practices, community impacts, and alternative governance models drawn from direct service to residents. Civil society, affected communities, and grassroots organizations should shape the agenda, not merely comment on it, through dedicated listening sessions, regional hearings, and reserved speaking roles. Technical experts, archivists, and human rights practitioners should jointly lead thematic tracks on archival integrity, data provenance, and safeguards against erasure. Structurally, the Dialogue should combine inclusive plenary sessions with small, problem focused working groups, publish all inputs and outputs openly, and establish ongoing thematic task forces rather than remaining a one off event.

Communities whose lives and histories are most shaped by data and state power remain underrepresented in global AI governance discussions. These include people of African descent and Indigenous communities whose histories have been fragmented or erased, migrants and undocumented people, residents of over policed neighborhoods, and communities living near environmental sacrifice zones. Grassroots disability justice organizers, low income workers subject to algorithmic management, and local archivists and memory workers are also rarely centered, despite their direct experience with opaque systems and disappearing records. Inclusion cannot mean only inviting them to react to pre set agendas. It requires resourcing community-led participation, funding interpretation and accessibility, and creating standing roles for these groups in advisory bodies, working groups, and agenda setting processes. It also means holding formal consultations in affected localities, not just in global capitals, and explicitly recognizing community-generated knowledge, oral histories, and grassroots documentation as legitimate expertise for shaping AI governance. Poets and other cultural workers should also be invited as contributors, drawing on methods like poetry to hold memory, grief, and resistance in forms policy language cannot reach. A global poetry slam alongside the Dialogue could offer a powerful venue for communities to narrate how AI touches their lives. As both researcher and poet, I would welcome the chance to help shape such spaces.

Innovative engagement formats should make the AI Dialogue participatory, creative, and accountable, rather than limited to expert panels and prepared statements. One approach is to pair every high level plenary with community testimony sessions where affected communities, local governments, and civil society present concrete cases of harm, resistance, and alternative practice. These could include multimedia storytelling, short films, and live demonstrations of how AI intersects with archives, policing, migration, or environmental decision making, followed by respondents required to propose specific governance actions, not only "take note." Second, transnational collaboratories and community "ecoversities" could bring together technical experts, archivists, poets, activists, and local officials to interrogate real systems and datasets, including those shaping historical records. Designed as ongoing learning and co creation spaces rather than one off workshops, they would produce tangible outputs such as draft standards, model safeguards, and new monitoring tools that can be tested and shared. Attribution and participation are critical here; these formats should include clear commitments that ideas, methods, and frameworks are credited to the communities and individuals who develop them, and that they are not replicated while excluding their originators. Creative and cultural formats, such as a global poetry slam, story circles, and community art exhibits, should be embedded as formal sessions. As both researcher and poet, I see these as critical venues for articulating memory, loss, and resistance where policy language falls short, and for ensuring that AI governance remains accountable to lived experience.

Several existing policies and practices offer concrete building blocks for effective AI governance that the AI Dialogue could learn from and extend. In Illinois, where I am from, the Biometric Information Privacy Act (BIPA) remains one of the strongest biometric privacy laws globally, requiring informed written consent, clear retention policies, and providing a private right of action against misuse of biometric data such as facial recognition and fingerprints. It shows how subnational law can meaningfully constrain harmful AI deployment. Illinois' AI in employment protections, including HB 3773 and related amendments to the Illinois Human Rights Act, require employers to disclose when AI is used in certain employment decisions and prohibit AI driven practices that result in unlawful discrimination, building on the Artificial Intelligence Video Interview Act, which already mandates notice, explanation, and consent for AI analyzed video interviews. State AI policy for public sector use, such as Illinois' Policy on the Acceptable and Responsible Use of Artificial Intelligence, sets rules for acceptable and responsible AI in state agencies, including constraints on how state data may be used, demonstrating that public sector AI must be governed by explicit standards. The Dialogue should also look to countries often left out of AI debates, such as Venezuela, Uruguay, and Kenya, where important emerging AI policy work is underway. In addition, the framing and language of the Hiroshima AI Process Friends Group offer a useful reference for multilateral, human centered AI governance that the Dialogue could further develop and globalize.