Deep‑time Digital Earth and GeoGPT claim geopolitical neutrality but cannot be independently verified: proprietary core code, concentrated funding and governance ties to a single national jurisdiction, and undisclosed content filtering mean the platform operates as a state‑anchored international service rather than a geopolitically neutral public good.

Main Finding

Using UNESCO Open Science and AI Ethics frameworks, the authors find that claims of geopolitical non‑alignment for Deep‑time Digital Earth (DDE) and GeoGPT cannot be independently verified. The platform is best characterised as a "state‑anchored international platform": globally participatory at the interface but technically and institutionally dependent on a single national jurisdiction. Core software and models remain proprietary; governance and funding are concentrated and linked to state‑linked initiatives, and AI content‑filtering practices are opaque and un‑auditable.

Key Points

• Claims vs verifiability: DDE/GeoGPT publicly claim geopolitical neutrality but fail the authors’ verifiability tests across governance independence, data/software openness, and AI transparency.
• Proprietary core: The platform’s essential software and models are not open‑source; the system diverges from UN/DPG/Open Science definitions of "openness" and Digital Public Goods standards.
• Governance concentration: Documentary evidence shows governance/funding concentrated in one national jurisdiction; GeoGPT governance shifted to Zhejiang Lab (co‑founded by Alibaba) in March 2025; DDE was suspended by the IUGS in early 2025 but neutrality claims persist.
• Geo‑political links: Authors document links between DDE collaboration targets and China’s Belt and Road Initiative, a 2026 agreement with the mining investment arm of the China‑Africa Development Fund, and advisory ties to actors with resource‑extraction affiliations.
• Opaque filtering: Governance documentation reveals content filtering of politically sensitive material (one instance replaced a peer‑review citation on state censorship with a refusal); the criteria/scope of filtering are undisclosed and un‑auditable.
• Transferable test: The three criteria (governance independence, data/software openness, AI transparency) operationalised from UNESCO/FAIR/ISC provide a transferable method to assess neutrality claims of comparable platforms.
• Practical recommendation: Given the verification gap, engagement should be calibrated; preference advised for open‑source/sovereign alternatives where data sovereignty, privacy, or economic leverage is at stake.

Data & Methods

• Case selection: Single‑case study—DDE and GeoGPT—chosen because they are the only identified international AI‑enabled geoscience platform explicitly claiming geopolitical non‑alignment.
• Analytical framework: Combines UNESCO Open Science and AI Ethics Recommendations, FAIR principles, ISC frontier AI guidance, CARE principles and Digital Public Goods standards to define three operational criteria: governance independence; data & software openness; AI transparency/explainability.
• Data collection: English and Chinese documentary searches for governance documents, MoUs, terms of use, platform agreements, academic and grey literature, and publicised partnerships/agreements (including 2026 China‑Africa Development Fund deal).
• Units of analysis: Governance structures and funding sources; openness of data, code, and deployability of models; disclosure of model training data, weights, parameters, and documented AI behaviours (e.g., content filtering).
• Interpretive approach: Socio‑technical institutional analysis emphasising that surface accessibility (open interface) can mask core technical opacity; conclusions drawn from cumulative pattern across unresolved verifiability cases.

Implications for AI Economics

• Market power & vendor lock‑in: Proprietary cores and concentrated governance raise risks of vendor lock‑in and potential extraction of rents by platform controllers, affecting bargaining positions of states and firms using the platform.
• Data sovereignty as economic leverage: Control over geological data and analytical platforms directly affects resource valuation, investor negotiation leverage, and sovereign revenue outcomes—opaque platforms can distort these economic signals.
• Investment and path dependence: Adoption of externally governed platforms can create path dependence (standards, formats, workflows) that shapes future investment, capabilities and switching costs for African institutions.
• Externalities & informational bias: Opaque training data/models and undocumented filtering may embed political or institutional biases that systematically affect resource estimates, risk assessments, and economic decision‑making—these are negative externalities with measurable economic consequences.
• Public good vs private capture: The case highlights the economic value of truly open public goods (auditability, reusability, competition) for lowering transaction costs, enabling local value capture, and fostering competitive markets for AI‑enabled geoscience services.
• Policy and procurement: Governments and multilateral funders should incorporate verifiability criteria into procurement, funding, and partnership requirements (e.g., demand open licensing, auditable models, multilateral governance) to protect economic sovereignty.
• Capacity building tradeoffs: Short‑term gains from adopting externally hosted tools must be weighed against long‑term costs of lost capability, decreased bargaining power, and dependence; economic policy should invest in sovereign compute, open‑source stacks, and local model development.
• Standardisation and regulation role: Multilateral standards (e.g., DPG, UNESCO/ISC guidance) and regulatory instruments can mitigate asymmetric power, but require enforceable conditions tied to funding and collaboration to be effective.

Short conclusion: For economic actors (governments, investors, firms) in Africa and elsewhere, reliance on AI‑enabled geoscience platforms that are not open and auditable creates quantifiable economic risks—policy should prioritise verifiability, sovereign alternatives, and open public goods to safeguard economic sovereignty and efficient markets.