This page is still being polished. If you have thoughts, please share them via the feedback form.
Data on this page is preliminary and may change. Please do not share or cite these figures publicly.
Structured analysis to identify, characterize, and prioritize potential harms and risks.
Also in Risk & Assurance
Comprehensive risk assessments before deployment that would assess reasonably foreseeable misuse and include dangerous capability evaluations that incorporate post-training enhancements and collaborations with domain experts. Risk assessments would inform deployment decisions.
Multiple experts agreed that pre-deployment risk assessments could be helpful across all risk areas as a general best practice. They noted that such assessments allow for proactive identification and mitigation of potential risks before AI models are deployed. Several experts emphasised the importance of involving domain experts in the assessment process to ensure thorough and informed evaluations. Some experts cautioned that the effectiveness of these assessments would depend heavily on who performs them and their level of independence. Several mentioned that internal assessments might be biased towards deployment due to conflicts of interest. Several experts pointed out that pre-deployment assessments may be more effective for certain risks (like chemical, biological, radiological, and nuclear (CBRN) risks; and critical infrastructure disruption) that are more concrete and easier to test for, compared to more diffuse risks like impacts on democratic processes. Several experts noted that while helpful, these assessments have limitations. They mentioned the difficulty in predicting all potential risks, especially for unforeseen or evolving threats. Some suggested that ongoing monitoring and iterative assessments would be necessary to complement pre-deployment evaluations. Individual experts raised additional points, such as the need for standardisation, the potential for false negatives, and the importance of these assessments being substantive rather than mere box-ticking exercises.
Reasoning
Structured analysis identifying and characterizing foreseeable harms before system deployment.
Third party pre-deployment model audits
External pre-deployment assessment to provide a judgment on the safety of a model. Auditors, which could be governments or independent third parties, would receive access to a fine-tuning API for testing, or further appropriate technical means.
2.2.3 Auditing & ComplianceExternal assessment of testing procedure
Bringing in external AI evaluation firms before deployment to assess and red-team the company's execution of dangerous capabilities evaluations.
2.2.2 Testing & EvaluationVetted researcher access
Giving good faith, public interest evaluation researchers access to black-box research APIs that provide technical and legal safe harbours to limit barriers imposed by usage policy enforcement, logging, and stringent terms of service.
2.3.1 Deployment ManagementAdvanced model access for vetted external researchers
Examples of advanced access rights could include any of the following: increased control over sampling, access to fine-tuning functionality, the ability to inspect and modify model internals, access to training data, or additional features like stable model versions.
2.2.2 Testing & EvaluationData curation
Careful data curation prior to all development stages (including fine-tuning) to filter out high-risk content and ensure the training data is sufficiently high-quality.
1.1.1 Training DataHarmlessness training
State-of-the-art reinforcement learning and fine-tuning techniques, such as Reinforcement Learning from Human Feedback (RLHF) or Direct Preference Optimization (DPO), to ensure models do not engage in unsafe behavior.
1.1.2 Learning ObjectivesEffective Mitigations for Systemic Risks from General-Purpose AI
Uuk, Risto; Brouwer, Annemieke; Schreier, Tim; Dreksler, Noemi; Pulignano, Valeria; Bommasani, Rishi (2024)
The systemic risks posed by general-purpose AI models are a growing concern, yet the effectiveness of mitigations remains underexplored. Previous research has proposed frameworks for risk mitigation, but has left gaps in our understanding of the perceived effectiveness of measures for mitigating systemic risks. Our study addresses this gap by evaluating how experts perceive different mitigations that aim to reduce the systemic risks of general-purpose AI models. We surveyed 76 experts whose expertise spans AI safety; critical infrastructure; democratic processes; chemical, biological, radiological, and nuclear risks (CBRN); and discrimination and bias. Among 27 mitigations identified through a literature review, we find that a broad range of risk mitigation measures are perceived as effective in reducing various systemic risks and technically feasible by domain experts. In particular, three mitigation measures stand out: safety incident reports and security information sharing, third-party pre-deployment model audits, and pre-deployment risk assessments. These measures show both the highest expert agreement ratings (>60\%) across all four risk areas and are most frequently selected in experts' preferred combinations of measures (>40\%). The surveyed experts highlighted that external scrutiny, proactive evaluation and transparency are key principles for effective mitigation of systemic risks. We provide policy recommendations for implementing the most promising measures, incorporating the qualitative contributions from experts. These insights should inform regulatory frameworks and industry practices for mitigating the systemic risks associated with general-purpose AI.
Plan and Design
Designing the AI system, defining requirements, and planning development
Developer
Entity that creates, trains, or modifies the AI system
Map
Identifying and documenting AI risks, contexts, and impacts
