Geopolitical Insight and Education Foundation

The UN Global Dialogue on AI Governance is the most consequential forum for collective, global action on the challenges confronting society due to the enormous rise of generative AI in all aspects of our economic, social and political life. Success would entail generated collective action on the many highly important issues surrounding AI Governance, and ensuring a broader understanding of the interconnection between these issues, and what underpins them. The position of the Geopolitical Insight and Education Foundation is that epistemic trust in the scientific record and society as a whole is one of the most fundamental bedrock shared institutions from which we can and must build a clear mutual understanding of our global society, and that is why it is our purpose to safeguard it from trust collapse due to AI usage in science.

• Transparency, accountability, and human oversight
• Safe, secure and trustworthy AI
• Social, economic, ethical, cultural, linguistic and technical implications of AI
• Interoperability of governance approaches

• The focus of the Geopolitical Insight & Education Foundation is on democratic stability and resilience. We believe that strong AI governance will be crucial to safeguarding our democracies and the conceptual institutions - like trust, factfulness, verifiability, and science - that underpin it. Thus, our Responsible AI in Science campaign directly promotes safe, secure and trustworthy AI
• exploration of the social, economic, ethical, cultural, linguistic and technical implications of AI
• the importance of interoperability of governance approaches through a standardized framework to AI governance in science
• and transparency, accountability and human oversight to ensure the safety of the scientific record and trust in our societies.

The scientific record is one of civilization's most consequential public institutions. It is the foundation on which all policy is built and all challenges are addressed, from medicine to climate policy to technological innovation to democratic governance. Science as an institution functions because it is cumulative, verifiable, and trusted. This epistemic trust in science is one of the most important foundations of stability in our democracies. That trust is now under existential threat. Generative AI is now entering the scientific record at a speed and volume that the self-correcting institutions science relies on cannot combat. Synthetic text, references, and data are being published, cited, and built upon, while disclosure is inconsistent and enforcement is weak. Once embedded, AI-generated material is reused and, after only a few cycles, becomes harder to audit. At that point, work grounded in real evidence and work built on synthetic science become indistinguishable. Further, the gatekeeping mechanism itself is now compromised by the same technology it is supposed to evaluate: Peer review is increasingly being automated by AI systems, meaning the system that validates knowledge is being flooded and compromised at the same time. This is already happening, and it is a threat to trust in the scientific record as a whole. Consequently, if we lose the ability to reliably distinguish what is true, we lose the basis on which our democratic societies operate.

There is a clear intervention point on the problem of AI entering the scientific record. A few major academic publishing conglomerates publish the vast majority of all published science. No other actors in the scientific ecosystem have the reach, the authority, or the structural position to act at the scale this moment requires. This is an enormous responsibility, but it is also an historic opportunity to be the protectors of the scientific record. Through our Responsible AI in Science program, we aim to work collaboratively with the major academic publishers, as well as policymakers, institutions, and experts to align detection, disclosure, and enforcement standards. We are calling for alignment of AI usage and disclosure requirements, movement beyond self-reporting toward real detection and real enforcement mechanisms, and the formation of a publisher-led coalition to build the governance infrastructure science now needs. We urge the Global Dialogue on AI Governance and its participants to recognize the integrity of the scientific record as an urgent priority for coordinated international action. Safeguarding epistemic trust is a condition for effective governance, for international cooperation, and for the stability of our societies.

This is a tractable problem, but it is time-sensitive. The scientific record took centuries to build. It will not take centuries to lose. We urge the Global Dialogue on AI Governance and its participants to recognize the integrity of the scientific record as an urgent priority for coordinated international action. Safeguarding epistemic trust is a condition for effective governance, for international cooperation, and for the stability of our societies.

Our Responsible AI in Science campaign, focused on safeguarding epistemic trust in society, by acting at the leverage point of the 12 major academic publishers that are the gatekeepers of what enters the scientific record. We are launching an Open Letter, convening policymakers, academics, scientists, AI developers, intergovernmental institutions and public society, as we aim to work collaboratively with the academic publishers on this issue.

We believe that, in addition to the impactful forum hosted here by the UN Global Dialogue on AI Governance, stakeholders can contribute their expertise through various civil society initiatives; one of which is including (but certainly not limited to) our Responsible AI in Science advocacy campaign and Open Letter.

Academics, scientists and research-producers globally do not yet have the voice that they need, to promote the transparent, open, clear, verifiable and trusted production of science, which serves as the bedrock of epistemic trust in society.

Convening different types of experts to engage in effective and multi-role exploratory dialogues could be useful. From our perspective, we try to convene a wide cross-section of society for our advocacy - from academics, to influencers/communicators, to policymakers, to companies, to media, to citizen initiatives, to international bodies. Representative voices from each make the whole stronger.

Our policy proposals for the Responsible AI in Science campaign are: 1) Align AI usage and disclosure requirements across all scientific journals, publishing portfolios and institutions into a global gold standard; 2) Move beyond self-reporting in science toward real detection and real enforcement mechanisms; 3) Forming a publisher-led coalition to build the governance infrastructure science now needs to safeguard epistemic trust in society.