Education became an agent deployment site
AI in education is often framed as a tutoring question. In 2025 it became a systems question: who sets the learning goal, who sees the student’
THE UAF INDEX
Policy
The open-model bargain had a bill
Open models can widen access to capability, lower experimentation costs and create alternatives to a small number of providers. They also move more responsibility to the
Paris widened the AI argument
The 2025 Paris AI Action Summit widened the public AI argument beyond frontier-model risk. Governments, companies, researchers and civil society discussed access, sustainability, public infrastructure
Open safety data needed a maintenance plan
Opening a dataset can broaden AI research, but openness is not the same as stewardship. Safety data needs provenance, documentation, versioning and a clear account of
Operator can click. You still have to mean it.
OpenAI’s browser agent can act on your behalf. Its own system card makes clear why handing over the clicks is not the same as handing over responsibility.
The system card became part of the launch
Operator’s system card demonstrated why a model release increasingly needs a public technical record. A product page describes what a system can do; a system
MCP offers a plug. You supply the permission.
Anthropic’s protocol gives AI tools a common way to connect. Deciding what those connections permit is still someone’s job.
The Machine Has a Policy. Who Checks the Paperwork?
Risk frameworks and company policies describe oversight. Recorded assessments and disclosed procedural lapses show why implementation still needs scrutiny.
Autonomous Does Not Mean Unaccountable
Autonomy describes room to act. The harder questions concern who grants it, who checks it and what changes when capability grows.
Attention is not a permission slip
The Transformer, GPT-3 and ReAct answer different technical questions. None of those answers should spare us from asking what a system may do.