On 4 November, the Global Network Initiative (GNI) released its 2025 Annual Report: Finding Resilience through Multistakeholderism, which includes, among other things, information about the company assessments that were conducted in that calendar year. Those assessments were the first in the fifth cycle of assessments since GNI’s inception in 2008, which incorporate several major changes in GNI’s assessment process.
In 2025, six member companies completed their assessments: Verizon, MTN, Wikimedia Foundation, Google, Meta, and Zoom. This was the first GNI assessment conducted for MTN, Wikimedia, and Zoom, while it constituted Meta’s fourth and Google’s fifth assessment. Company assessments conducted in 2026, including for Cloudflare, Ericsson, Microsoft, Nokia, and Ericsson, will be covered in the 2026 Annual Report.
This blog covers four key themes that have emerged so far from the assessments conducted in 2025.
The fifth-cycle assessment reports highlight the ways that the GNI Principles continue to be embedded into corporate governance.
Established members have, over multiple GNI assessment and recommendation cycles, operationalized the GNI Principles directly within their core corporate governance structures. Verizon implements its GNI commitments through its dedicated Business & Human Rights Program (BHRP) of attorneys who collaborate cross-functionally to assess product development risks, while also standardizing international operations via its “International Law Enforcement Request Guidelines” to mandate specialized human rights legal reviews for high-risk demands. Similarly, Google anchors its implementation efforts at the highest level by integrating its Human Rights Executive Council (HREC) into its formalized Trust & Compliance Council (TCC), and has grown its Human Rights Team to facilitate context-aware regional capacity and stakeholder engagement across global markets. Finally, Meta has established a joint executive oversight structure between its President of Global Affairs and Chief Legal Officer, with formal escalation pathways up to the CEO, while its Human Rights Team has embedded a comprehensive HRDD framework that proactively monitors human rights risks and uses formalized thresholds to trigger ongoing assessments.
Meanwhile, new member companies are leveraging preexisting structures to implement the GNI Principles, while simultaneously using the GNI framework to further enhance those governance frameworks. Zoom was commended for successfully integrating its corporate “Care” value into a robust Compliance and Trust & Safety framework with systematic government request handling, and was recommended to build on this foundation by establishing a dedicated, senior-directed human rights function to oversee its commitments. MTN was recognised for its strong foundational commitment to digital rights and the practical design of its Digital Human Rights Playbook, and was recommended to simplify and localize this playbook for non-expert, regional operating staff. Finally, the Wikimedia Foundation, which successfully relies on its decentralized, community-governed model to uphold the GNI Principles, was recommended to develop standardized, documented protocols to maintain consistency when responding to diverse government demands across international jurisdictions.
Synergies between the GNI framework and mandatory EU HRDD frameworks were explored in the fifth cycle. Several companies used their assessments to explain how they approach compliance with these regulations, noting where those steps have helped reinforce their GNI commitments, while others reported on how systems and policies they developed to implement the GNI Principles helped prepare them to demonstrate compliance with those regulations.
For example, Google streamlined its DSA compliance by integrating the regulation’s requirements directly into its established risk assessment and enterprise risk management processes, which have been informed over time by the GNI Principles. Meanwhile, the Wikimedia Foundation grounded its GNI assessment, as well as its inaugural DSA systemic risk assessment, in its global 2020 HRIA and 2024 CRIA registries. Both of these examples validate the work GNI has done over the last five years to advocate for synergies between voluntary and mandatory risk management and due diligence frameworks, and demonstrate how the incorporation of voluntary commitments through multistakeholder processes can help inform and facilitate regulatory compliance.
The 2025 assessments also provided insights on how companies are conducting AI-related due diligence and responding to AI-related government interventions in line with the GNI Principles. For example, as part of its GNI implementation, Zoom utilizes Privacy Impact Assessments for AI features and has revised its Acceptable Use Guidelines to explicitly clarify that customer content is not used to train AI models. Verizon’s BHRP collaborates with its Responsible AI team to address risks through specific deployment guidelines, while MTN identifies technical AI risks during product development via a partnership between its Legal, Sustainability, and Group Tech teams.
Google was commended for its rigor in Responsible AI development, with assessors recommending that the company sustain its investment in generative AI governance, while Meta’s assessor recognized its successful integration of the GNI Principles, including vis-a-vis Meta AI, across its global operations through robust policies and strategic engagement. In one case study, which has been anonymized in the annual report, a company highlighted how it handled an informal government request, including by making adjustments to its AI chat product to avoid hallucination in scam reporting.
The 2025 assessments also highlight the implementation of the GNI Principles during global crises. For instance, during the Russian invasion of Ukraine, a company disabled real-time location features, such as traffic heat maps and area busy-ness data, to prevent the targeting of civilians or troops while maintaining essential navigation tools to assist with evacuations. In the Israel/Palestine conflict, Meta adjusted its Dangerous Organizations and Individuals policy to narrow its scope, distinguishing between the “glorification” of designated entities and simple “praise” in order to prevent overmoderation. Additionally, the Wikimedia Foundation refused a demand from an authoritarian government to alter or remove content regarding a recent military conflict, choosing to face penalties rather than compromise the accuracy of important content.
The Path Forward
Although the fifth assessment cycle is still ongoing, GNI is already looking toward the future. Preparations for the sixth assessment cycle began in October 2025 following the Board’s approval of a strategic, phased “assessment evolution” workplan targeted for completion by the end of 2026. This phased approach will involve gathering member and accredited assessor feedback to maximize the value that assessments provide to members and to the tech ecosystem. Further details regarding the evolution of this process will be shared next year, including in the 2026 Annual Report, which is expected shortly in Q1 2027. Stay tuned!