AI Procurement Advisory — Tertiary Education Institution, Hong Kong
Service:
AI Procurement Advisory
Client:
Education Institution
Duration:
1 month
Date:
What AI Risks did the enterprise face?
LexGuard led the procurement review and contract negotiation, focusing on intellectual property, data-use permissions, confidentiality, and the treatment of institution-specific datasets and AI functionality.
How did LexGuard AI safeguard the client’s IP?
The proposed agreement did not adequately distinguish between:
The vendor’s pre-existing chatbot platform and underlying architecture;
The institution’s proprietary source data and structured datasets;
Data generated through the institution’s curriculum, administrative rules, and internal knowledge; and
AI models, algorithms, and logic flows developed specifically for the institution.
The terms also created a risk that the vendor could reuse, modify, license, or incorporate institutional data into products serving other customers.
The Results
IP boundaries were clarified. The vendor retained ownership of its underlying chatbot framework, while the institution retained ownership of its proprietary source data and structured datasets.
Data use was restricted. The vendor’s rights to reuse, modify, license, or use the institution’s data for broader model improvement were removed.
Model use was isolated. The AI model trained on the institution’s data could be used only for the institution’s benefit.
Institution-specific logic was protected. Algorithms and logic flows developed from the institution’s curriculum and administrative rules were classified as work-for-hire and assigned exclusively to the institution.
Confidentiality protections were strengthened. The final agreement prohibited the ingestion of student data and research IP into public or shared models.
The institution secured ownership of its fine-tuned models and structured data while preserving the vendor’s ownership of its pre-existing platform architecture. The revised agreement created a more defensible basis for deploying the AI solution without allowing the institution’s data or operational knowledge to become an unrestricted input into the vendor’s commercial product.
Testimonial
Lexguard's guidance gave us a clearer, more defensible basis for deploying AI while protecting student data and institutional intellectual property.
