Towards evaluations-based safety cases for AI scheming

• URL: http://arxiv.org/abs/2411.03336v2
• Date: Thu, 07 Nov 2024 09:18:26 GMT
• Title: Towards evaluations-based safety cases for AI scheming
• Authors: Mikita Balesni, Marius Hobbhahn, David Lindner, Alexander Meinke, Tomek Korbak, Joshua Clymer, Buck Shlegeris, Jérémy Scheurer, Charlotte Stix, Rusheb Shah, Nicholas Goldowsky-Dill, Dan Braun, Bilal Chughtai, Owain Evans, Daniel Kokotajlo, Lucius Bushnaq,
• Abstract summary: We propose three arguments that safety cases could use in relation to scheming. First, developers of frontier AI systems could argue that AI systems are not capable of scheming. Second, one could argue that AI systems are not capable of posing harm through scheming. Third, one could argue that control measures around the AI systems would prevent unacceptable outcomes even if the AI systems intentionally attempted to subvert them.
• Score: 37.399946932069746
• License:
• Abstract: We sketch how developers of frontier AI systems could construct a structured rationale -- a 'safety case' -- that an AI system is unlikely to cause catastrophic outcomes through scheming. Scheming is a potential threat model where AI systems could pursue misaligned goals covertly, hiding their true capabilities and objectives. In this report, we propose three arguments that safety cases could use in relation to scheming. For each argument we sketch how evidence could be gathered from empirical evaluations, and what assumptions would need to be met to provide strong assurance. First, developers of frontier AI systems could argue that AI systems are not capable of scheming (Scheming Inability). Second, one could argue that AI systems are not capable of posing harm through scheming (Harm Inability). Third, one could argue that control measures around the AI systems would prevent unacceptable outcomes even if the AI systems intentionally attempted to subvert them (Harm Control). Additionally, we discuss how safety cases might be supported by evidence that an AI system is reasonably aligned with its developers (Alignment). Finally, we point out that many of the assumptions required to make these safety arguments have not been confidently satisfied to date and require making progress on multiple open research problems.

Related papers

• Imagining and building wise machines: The centrality of AI metacognition [78.76893632793497]
  We argue that shortcomings stem from one overarching failure: AI systems lack wisdom. While AI research has focused on task-level strategies, metacognition is underdeveloped in AI systems. We propose that integrating metacognitive capabilities into AI systems is crucial for enhancing their robustness, explainability, cooperation, and safety.
  arXiv  Detail & Related papers  (2024-11-04T18:10:10Z)
• Taking AI Welfare Seriously [0.5617572524191751]
  We argue that there is a realistic possibility that some AI systems will be conscious and/or robustly agentic in the near future. It is an issue for the near future, and AI companies and other actors have a responsibility to start taking it seriously.
  arXiv  Detail & Related papers  (2024-11-04T17:57:57Z)
• Using AI Alignment Theory to understand the potential pitfalls of regulatory frameworks [55.2480439325792]
  This paper critically examines the European Union's Artificial Intelligence Act (EU AI Act) Uses insights from Alignment Theory (AT) research, which focuses on the potential pitfalls of technical alignment in Artificial Intelligence. As we apply these concepts to the EU AI Act, we uncover potential vulnerabilities and areas for improvement in the regulation.
  arXiv  Detail & Related papers  (2024-10-10T17:38:38Z)
• Towards Guaranteed Safe AI: A Framework for Ensuring Robust and Reliable AI Systems [88.80306881112313]
  We will introduce and define a family of approaches to AI safety, which we will refer to as guaranteed safe (GS) AI. The core feature of these approaches is that they aim to produce AI systems which are equipped with high-assurance quantitative safety guarantees. We outline a number of approaches for creating each of these three core components, describe the main technical challenges, and suggest a number of potential solutions to them.
  arXiv  Detail & Related papers  (2024-05-10T17:38:32Z)
• Safety Cases: How to Justify the Safety of Advanced AI Systems [5.097102520834254]
  As AI systems become more advanced, companies and regulators will make difficult decisions about whether it is safe to train and deploy them. We propose a framework for organizing a safety case and discuss four categories of arguments to justify safety. We evaluate concrete examples of arguments in each category and outline how arguments could be combined to justify that AI systems are safe to deploy.
  arXiv  Detail & Related papers  (2024-03-15T16:53:13Z)
• Artificial Intelligence: Arguments for Catastrophic Risk [0.0]
  We review two influential arguments purporting to show how AI could pose catastrophic risks. The first argument -- the Problem of Power-Seeking -- claims that advanced AI systems are likely to engage in dangerous power-seeking behavior. The second argument claims that the development of human-level AI will unlock rapid further progress.
  arXiv  Detail & Related papers  (2024-01-27T19:34:13Z)
• Managing extreme AI risks amid rapid progress [171.05448842016125]
  We describe risks that include large-scale social harms, malicious uses, and irreversible loss of human control over autonomous AI systems. There is a lack of consensus about how exactly such risks arise, and how to manage them. Present governance initiatives lack the mechanisms and institutions to prevent misuse and recklessness, and barely address autonomous systems.
  arXiv  Detail & Related papers  (2023-10-26T17:59:06Z)
• AI Deception: A Survey of Examples, Risks, and Potential Solutions [20.84424818447696]
  This paper argues that a range of current AI systems have learned how to deceive humans. We define deception as the systematic inducement of false beliefs in the pursuit of some outcome other than the truth.
  arXiv  Detail & Related papers  (2023-08-28T17:59:35Z)
• Cybertrust: From Explainable to Actionable and Interpretable AI (AI2) [58.981120701284816]
  Actionable and Interpretable AI (AI2) will incorporate explicit quantifications and visualizations of user confidence in AI recommendations. It will allow examining and testing of AI system predictions to establish a basis for trust in the systems' decision making.
  arXiv  Detail & Related papers  (2022-01-26T18:53:09Z)
• AI Failures: A Review of Underlying Issues [0.0]
  We focus on AI failures on account of flaws in conceptualization, design and deployment. We find that AI systems fail on account of omission and commission errors in the design of the AI system. An AI system is quite likely to fail in situations where, in effect, it is called upon to deliver moral judgments.
  arXiv  Detail & Related papers  (2020-07-18T15:31:29Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.