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.
Input validation, output filtering, and content moderation classifiers.
Also in Non-Model
Monitoring for dangerous outputs (e.g., code that appears to be malware or viral genome sequences) and inputs that violate acceptable use policies to ensure models do not engage in harmful behavior.
Multiple experts noted that input/output filtering could be effective for specific risks like chemical, biological, radiological, and nuclear (CBRN) risks; and cybersecurity issues, where harmful content is more clearly definable. However, several also pointed out that sophisticated actors might still find ways to circumvent these filters. Many experts expressed scepticism about the effectiveness of filtering for more complex or nuanced risks like bias, discrimination, and threats to democratic processes. They argued these issues are harder to define and detect through simple filtering mechanisms. Several experts mentioned that while filtering can be useful, it has limitations. Some noted it may be more effective for average users but less so for determined malicious actors. Others pointed out that overzealous filtering could potentially infringe on free speech or hinder legitimate uses. Several experts highlighted the importance of monitoring inputs and outputs for identifying misuse patterns and improving risk mitigation strategies over time. However, concerns about privacy and data protection were also raised. Some experts emphasised the need for clear definitions of harmful content and transparent policies governing the filtering process. A couple mentioned the challenge of keeping filters up-to-date with evolving threats.
Reasoning
Filters inputs/outputs to block dangerous content before model engagement or user delivery.
Pre-deployment risk assessments
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.
2.2.1 Risk AssessmentThird 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 DataEffective 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.
Operate and Monitor
Running, maintaining, and monitoring the AI system post-deployment
Deployer
Entity that integrates and deploys the AI system for end users
Measure
Quantifying, testing, and monitoring identified AI risks
