UNESCO and experts warn universities against technological lock-in with AI systems
At Digital Learning Week 2026, UNESCO called for auditability of AI in education, data portability, and open systems. Example: UNSW Sydney deployed ChatGPT Edu for over 80 000 students and staff, while also building a multi-model environment so as not to become dependent on a single vendor.
At the UNESCO Digital Learning Week event in Paris in September 2026, education ministers called for AI systems used in education to be auditable, for data from them to be portable, and for institutions to consider total cost of ownership before deployment. According to Tech Policy Press, the resulting ministerial declaration prioritizes interoperability, portability, and open-source solutions so that educational institutions retain the ability to switch vendors.
The text describes, as an example of the risk, universities gradually building an entire architecture around a single AI vendor: staff get used to the platform, students learn through it, their own assistants are created, and connections to other systems emerge. In such a situation, switching vendors no longer means simply replacing a software license, but untangling an entire ecosystem — universities have already experienced a similar problem with learning management systems, cloud infrastructure, or academic databases.
A specific example given is UNSW Sydney, which deployed ChatGPT Edu for more than 80 000 students and staff — one of the largest deployments of this kind in the Asia-Pacific region. According to the article, the university is at the same time building a multi-model environment in which various AI systems can be tested, in order to preserve the ability to choose between vendors. The author recommends that universities separate the institutional layer (identity, data access, security, usage rules) from the specific model running beneath it. Also mentioned is Will Stronge of the Autonomy Institute, who refers to AI systems in education as "cognitive infrastructure" and ranks questions of ownership and control over it among the key topics of education policy. The rest of the article was not available.
Why it matters
This is a warning about a risk that concerns not only universities but institutions and companies in general deploying commercial AI platforms into key processes: the more deeply AI becomes interwoven with other systems and workflows, the more expensive and complicated it later becomes to switch to a different provider. UNESCO's recommendations — auditability, data portability, open standards — offer concrete criteria by which contracts with AI vendors can be assessed already at the selection stage, not only when attempting to leave.
Relevant practical impact
What this means
For a business
Institutions and companies considering deploying a commercial AI platform into broader processes should, already at the procurement stage, assess auditability, data portability, and the ability to separate their own institutional layer (identity, data access, rules) from the specific model, in order to avoid costly lock-in with a single vendor.
Risks and complianceCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.