Summary
The paper looks at how to control who gets to use advanced AI and how powerful that AI is. Instead of just hiding AI behind an app or sharing the full system openly, the authors suggest a tiered system with three levels. Each level matches how capable the AI is and who can use it safely, from basic learning tools to fully powerful AI under strict rules. They explain why just sharing all AI details is risky and how safety measures can be undone if AI weights are open. To keep AI safe, they argue it’s not enough to limit AI power; access must be controlled by rules, too.
What this means in practice
- •For ai governance teams: Implement tiered access controls to match AI system capabilities and reduce misuse risks in organizations.
- •For cloud service managers: Enforce structured and accountable access to powerful AI models to comply with safety recommendations and regulatory frameworks.
A position paper. It proposes an approach and reports no results.
Abstract
Most debate about releasing capable AI systems treats the choice as binary: keep the model behind an API, or publish the weights. That framing runs together two separate decisions. One is how capable a system may become. The other is who may obtain a given capability, and on what terms. This paper is about the second decision. We propose a Multi-Tiered AI Accessibility Model (MAIAM) that sorts access by purpose, safety, and capability instead of by release artifact. Tier 1, AI for Learning, is broadly accessible and deliberately capability-bounded. Tier 2, AI for Exploration, serves research and development under registration, identity, and usage controls that tighten as capability rises. Tier 3, Full-Scale Frontier AI, is available only through structured, recorded, and accountable access. The principle behind the tiers is that accessibility should be proportional to capability, safety, and risk. We show why publishing frontier-class weights forfeits every safety lever a developer has after release, and, using evidence that safety fine-tuning can be stripped from open weights for a few hundred dollars, why an evaluation of the base model says little about its derivatives. We give criteria for assigning a system to a tier along capability, autonomy, domain-consequence, and usage/safety axes, and combine MAIAM with the SAFE AI Operating Point into a two-part framework: a Capability Boundary that fixes the safety limits not to be crossed, and an Accessibility Boundary that fixes who may obtain a capability below that limit. Capability limits alone do not make AI safe; a tiered access regime administered by regulating bodies is also needed.