Structured access: an emerging paradigm for safe AI deployment

By Toby Shevlane. 2022

62 citations recorded by the source corpus.

Structured access is an emerging paradigm for the safe deployment of artificial intelligence (AI). Instead of openly disseminating AI systems, developers facilitate controlled, arm's length interactions with their AI systems. The aim is to prevent dangerous AI capabilities from being widely accessible, whilst preserving access to AI capabilities that can be used safely. The developer must both restrict how the AI system can be used, and prevent the user from circumventing these restrictions through modification or reverse engineering of the AI system. Structured access is most effective when implemented through cloud-based AI services, rather than disseminating AI software that runs locally on users' hardware. Cloud-based interfaces provide the AI developer greater scope for controlling how the AI system is used, and for protecting against unauthorized modifications to the system's design. This chapter expands the discussion of"publication norms"in the AI community, which to date has focused on the question of how the informational content of AI research projects should be disseminated (e.g., code and models). Although this is an important question, there are limits to what can be achieved through the control of information flows. Structured access views AI software not only as information that can be shared but also as a tool with which users can have arm's length interactions. There are early examples of structured access being practiced by AI developers, but there is much room for further development, both in the functionality of cloud-based interfaces and in the wider institutional framework.

Subfields: Artificial Intelligence, Information Systems, Safety Research, Safety, Robustness & Interpretability, AI Ethics, Fairness & Society

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

1. Dan Hendrycks
2. Mantas Mazeika
3. Percy Liang
4. Steven Basart
5. Luke S. Zettlemoyer
6. Quoc V. Le
7. Dawn Song
8. Jianfeng Gao

In Information Systems

1. Irene Solaiman
2. Mantas Mazeika
3. Ponnurangam Kumaraguru
4. Jimmy Ba
5. Dan Hendrycks
6. Lorrie Faith Cranor

In Safety Research

1. Irene Solaiman

In Safety, Robustness & Interpretability

1. Mantas Mazeika
2. Dan Hendrycks
3. Steven Basart
4. Andy Zou
5. Alexander Pan
6. Irene Solaiman
7. Ann-Kathrin Dombrowski
8. Igor Mordatch

In AI Ethics, Fairness & Society

1. Irene Solaiman
2. Dan Hendrycks

Citations in the corpus

• The Gradient of Generative AI Release: Methods and Considerations (2023, methodology): “This method is favorable for structured access where research is possible but can still be tracked and is unlikely to create a “modified version” [86].”
• The WMDP Benchmark: Measuring and Reducing Malicious Use With Unlearning (2024, methodology): “In these cases, unlearning can be paired with structured API access (Shevlane, 2022), where model developers serve the unlearned model to everyday users, but serve the unrestricted, base model to approved users, such as red-teamers, security professionals, or virology researchers (Section 6.2).”